Abstract Peripheral T-cell lymphomas (PTCL) are aggressive mature T-cell malignancies associated with poor responses to standard therapies and frequent relapse, even in the era of targeted treatments. These outcomes highlight the critical need for biology-driven therapeutic strategies. PTCL is characterized by a complex tumor microenvironment (TME), in which malignant T-cells often represent only a minor fraction within a diverse immune and stromal ecosystem. While prior studies have largely focused on tumor-intrinsic features, the role of TME in driving lymphoma progression and therapeutic resistance remains largely unknown. Using a murine Vav1-myo1f model, we previously demonstrated that malignant T-cells actively shape an M2 macrophage-rich TME, and that targeting tumor-associated macrophages (TAMs) yields significant anti-lymphoma effects, supporting a critical role for the TME in PTCL pathogenesis. Here, we performed an integrated multiomic analysis of 29 human PTCL biopsies, including T follicular helper lymphoma (TFHL) and PTCL-NOS, and six reactive lymphoid controls using single-nucleus RNA sequencing, bulk RNA sequencing, and targeted mutational profiling. After quality control, 230,174 cells were analyzed. Unsupervised clustering identified 45 transcriptionally distinct TME cell states spanning benign T-cell, B-cell, myeloid, endothelial, and stromal compartments, many not previously described in PTCL. Tumors exhibited extensive remodeling of both cellular composition and transcriptional programs, including depletion of antigen-presenting macrophages, expansion of TAM states, enrichment of exhausted CD8+ T cells, and profound stromal reprogramming. Systems-level analysis identified recurrent multicellular ecosystems defined by coordinated gene expression programs across immune and stromal compartments. These ecosystems were associated with distinct cytokine signaling networks and included two mutually exclusive TME states: an inflammatory macrophage/FDC-associated ecosystem and an immunosuppressive M2 macrophage/cancer-associated fibroblast (CAF) ecosystem driven by TGFβ signaling. Together, these data demonstrate that PTCL is characterized not only by altered lineage abundance but also by reproducible remodeling of intra-lineage transcriptional programs, including the emergence of tumor-specific cellular states within the microenvironment. These structured ecosystems reflect coordinated cross-talk between malignant T-cells and surrounding immune and stromal populations and are shaped by distinct cytokine signaling circuits. Ligand-receptor analysis identified TGFβ-TGFBR interactions as central mediators of lymphoma-TME communication, and functional validation in a murine model showed that TGFβ blockade significantly reduced tumor burden and proliferation. Collectively, these findings provide new insights into PTCL biology and establish a foundation for TME-informed therapeutic strategies, with potential implications for patient stratification and targeted intervention. Citation Format: Wen-Hsuan Wendy Lin, Anqi Wang, Jean-Baptiste Reynier, Bobby B. Shih, Laura Quevedo, Craig R. Soderquist, Ryan Najac, Jianhua Wang, Hannah Miller, Cindy Ma, Anouchka P. Laurent, Ruth Alonso Alonso, Alyssa Bouska, Javeed Iqbal, Govind Bhagat, Adolfo Ferrando, Raul Rabadan, Teresa Palomero. Microenvironmental ecosystems in peripheral T-cell lymphomas reveal therapeutic vulnerabilities [abstract]. In: Proceedings of the Fifth AACR International Meeting on Advances in Malignant Lymphoma: From Discovery to Clinical Impact; 2026 Jun 24-27; Philadelphia, PA. Philadelphia (PA): AACR; Blood Cancer Discov 2026;7(3_Suppl):Abstract nr A056.

Mature T-cell lymphomas (TCL) are a heterogeneous group of non-Hodgkin lymphomas arising from the malignant transformation of post-thymic T and NK cells. These lymphomas are characterized by poor response to chemotherapy, high relapse rates, and dismal survival outcomes. Genomic analyses of TCL have revealed recurrent alterations affecting intracellular signaling pathways and epigenetic regulators. The frequent involvement of epigenetic modifiers suggests that epigenetic dysregulation is a critical step in TCL pathogenesis. To identify the epigenetic changes driving TCL transformation, we analyzed chromatin modifications and gene expression profiles in murine TCL models that closely recapitulate human disease. Our data revealed that Id2, a repressor of basic helix-loop-helix proteins, is consistently overexpressed at both RNA and protein levels across multiple murine lymphoma models. Supporting this, analysis of the Cancer Cell Encyclopedia identified TCL as one of the tumor types with the highest ID2 expression, which we confirmed by Western blot in human TCL-derived cell lines. Furthermore, analysis of single-nuclei RNAseq data from a cohort of primary angioimmunoblastic T-cell lymphoma (AITL) and peripheral T-cell lymphoma not otherwise specified (PTCL, NOS) patient samples generated in our laboratory, demonstrated significant ID2 overexpression in a subset of cases (12/27; 44%) compared to tonsil and lymph node controls. Interestingly, all relapsed TCL cases in our cohort showed significant overexpression of ID2. Based on our data, we hypothesize that ID2 is epigenetically regulated during transformation and contributes to T-cell lymphomagenesis. To define the mechanisms leading to ID2 overexpression in TCL, we performed H3K27ac and H3K4me1 ChIP-seq, ATAC-seq, reverse-ChIP coupled with mass spectrometry, and dual-luciferase reporter assays. These analyses identified a conserved region near the ID2 locus enriched for H3K27ac and H3K4me1 marks that co-localizes with ATAC-seq peaks and is ranked as a super-enhancer by the ROSE algorithm. This region is specifically activated in mature TCL cells compared to control CD4⁺ T cells. Reporter assays using the most conserved segment within the enhancer region showed more than twofold increased luciferase activity relative to a minimal promoter, confirming its enhancer function. Interestingly, reverse-ChIP identified CTCF and ASCL1 as candidate regulators of enhancer looping and transcriptional activation, respectively. To characterize the proteins mediating the epigenetic function of ID2 in TCL, we employed a biotin ligase proximity labeling assay (BioID) approach coupled with mass spectrometry. Our results identified known ID2 partners, including the E-proteins TCF3, TCF4 and TCF12, key regulators of T-cell differentiation, as well as novel partners including elements of the SWI/SNF chromatin-remodeling complex, suggesting that ID2 overexpression indirectly regulates transcriptional programs critical for lymphomagenesis. To determine the functional role of ID2 in TCL, we characterized the effects of both genetic and pharmacological inhibition of ID2, using shRNA and CRISPR-Cas9 mediated knockout and novel pan-ID inhibitors AGX51 and AGXA, respectively. Loss of ID2 leads to decreased cell viability and increased apoptosis in in vitro TCL models. Interestingly, RNA-seq analysis revealed enrichment in signatures associated with glucose metabolism, suggesting a novel role for ID2 in tumor metabolic regulation. Finally, in vivo treatment with the pan-ID inhibitor AGX51 in a murine lymphoma model induced a modest but consistent reduction in tumor burden, spleen weight, and activated CD69⁺ T cells population in AGX51-treated mice compared to vehicle-treated controls. In summary, we have identified a novel TCL-specific super-enhancer that drives ID2 overexpression in TCL and demonstrated an essential role for ID2 in supporting lymphoma proliferation and survival. More importantly, pharmacological targeting of ID2 reduces tumor burden in vitro and in vivo, highlighting its potential as a therapeutic target in mature T-cell lymphomas.

Angioimmunoblastic T-cell lymphoma (AITL) is an aggressive subtype of peripheral T-cell lymphoma (PTCL) derived from malignant transformation of T follicular helper (TFH) cells. AITL, which represents 20–30% of PTCLs, is associated with autoimmune features, poor response to conventional chemotherapy, and dismal prognosis. The molecular landscape of AITL is characterized by frequent genomic alterations in epigenetic regulators, including TET2, DNMT3A, and IDH2, as well as components of the T-cell receptor (TCR) signaling pathway. Among the latter, the RHOA G17V mutation has been recognized as a genetic hallmark uniquely associated with TFH-derived lymphomas. Previous work from our group demonstrated that the RHOA G17V mutation induces TFH specification and, in cooperation with the loss of Tet2, promotes lymphomagenesis. Recent data from our laboratory has shown that expression of RHOA G17V regulates the sphingosine-1-phosphate (S1P) receptor 1 (S1PR1), a key mediator of lymphocyte trafficking and immune activation. Expression of RHOA G17V in CD4⁺ T cells antagonizes S1PR1 internalization, leading to increased activation of S1PR1 downstream signaling. Interestingly, deletion of S1pr1 in CD4⁺ T cells in a murine Rhoa G17V conditional model partially rescues the thymic phenotype and peripheral systemic inflammation characteristic of RHOA G17V-expressing mice, suggesting a contribution of S1PR1 to the aberrant function of RHOA G17V. This, together with our observation of increased S1PR1 expression in human RHOA G17V⁺ AITL tumor samples, identifies S1PR1 deregulation as a potentially relevant effector of the oncogenic effects of RHOA G17V in TFH cells. S1P receptor inhibitors, such as fingolimod (FTY720) and next-generation, more selective modulators such as ozanimod, siponimod, and ponesimod, are FDA-approved for the treatment of multiple sclerosis and other autoimmune diseases. These agents function by inhibiting S1P receptors, thereby sequestering lymphocytes in lymphoid tissues and reducing pathogenic immune cell trafficking. Their established safety and immunomodulatory effects make them attractive candidates for repurposing as targeted therapies in AITL, where S1PR1 signaling seems to be implicated in disease pathogenesis. However, their mechanistic effects in T-cell lymphoma cells remain to be characterized. To dissect the effects of S1PR1 inhibition, we used fingolimod (a pan-S1PR modulator) and ozanimod (a more specific S1PR1 and S1PR5 modulator), in murine preclinical models of AITL in vitro. Both compounds significantly reduced viability and induced apoptosis in the murine AITL cell line. In parallel, migration assays with ex vivo primary murine lymphoma cells revealed potent inhibition of both spontaneous and S1P-induced migration. We performed RNAseq in a Tet2⁻/⁻ RHOA G17V-expressing murine lymphoma line treated with fingolimod or ozanimod for 24 h. GSEA analysis revealed robust downregulation of IL6-JAK-STAT3, TNFA-NFκB, and apoptosis pathways (FDR < 0.25), supporting the existence of a proinflammatory transcriptional program sustained by S1PR1 signaling. To functionally validate the contribution of STAT3 to this circuit, we treated cells with stattic, a selective inhibitor of STAT3 phosphorylation. Stattic partially recapitulated the anti-lymphoma effects of fingolimod, suppressing migration and inducing apoptosis in vitro. Importantly, in vivo treatment with either fingolimod or stattic led to reduced tumor burden and systemic dissemination and was associated with decreased STAT3 phosphorylation. These findings define an S1PR1–STAT3 inflammatory loop that promotes survival and dissemination in RHOA G17V-driven AITL and identify this axis as a therapeutically actionable vulnerability. In summary, our preliminary data establish a mechanistic link between oncogenic RHOA signaling and S1PR1-mediated membrane receptor signaling, highlighting a previously unrecognized pathway in AITL pathogenesis. The demonstration that clinically approved S1P receptor modulators suppress lymphoma growth and dissemination provides a strong rationale for repurposing these agents as targeted therapies for AITL, encouraging future translational studies and clinical investigation.

Background: Valchlor (mechlorethamine gel) is a topical chemotherapy used to treat cutaneous T-cell lymphoma (CTCL). Despite its well-established role in CTCL treatment, limited data exist on its real-world use and cost trends across different provider types and geographic regions. Methods: We examined Centers for Medicare and Medicaid Services Medicare Part D prescriber-level data from 2018-2023 to identify trends in Valchlor prescribing. We analyzed total claims, total drug costs, and average cost per claim. Comparisons were made among provider specialties (e.g., dermatology vs. hematology-oncology vs. physician assistant) and across U.S. states. Statistical tests, including t-tests and ANOVA, were used to evaluate differences in cost per claim among subgroups. Results: In 2023, only 44 individual prescribers met the Medicare Part D public reporting threshold by submitting >10 Valchlor claims. In 2023, the top five states with the highest number of Valchlor prescriptions were New York (161 claims), California (104), Missouri (71), Ohio (62), and Connecticut (52). The top five prescribing cities were New York, NY (90 claims), Fairport, NY (59), New Haven, CT (52), Saint Louis, MO (46), and Chicago, IL (41). From 2018-2023, 3,945 Valchlor prescriptions were submitted to Medicare Part D. Dermatologists accounted for 3,113 (79% of total) claims, with an average cost per claim of ~$7,220. Physician assistants submitted 291 (7%) claims, averaging $6,097 per claim, while hematology-oncology specialists submitted 250 (6%) claims, with the highest average cost per claim at $9,346. Dermatologists, on average, wrote more Valchlor claims per prescriber than hematologist-oncologists (20.4 vs 12.1 claims per prescriber). Overall, the number of claims increased by 8.2%/year on average, and the cost per claim rose by 5.1% annually, indicating both growing utilization and rising cost. We further evaluated cost disparity by geographic distribution. States with the highest average cost per claim included Michigan ($17,193) and New Jersey ($13,740), while Indiana ($3,955) and Oregon ($4,184) had the lowest. A one-way ANOVA confirmed that these geographic differences were statistically significant (p < 0.001). These findings emphasize notable variation in Valchlor pricing and reimbursement across regions. There was no correlation between the cost per claim of Valchlor and the median household income of the state. There was also no relationship between the number of claims in a state and that state's average cost per claim. Almost all (99.4%) claims originated in a state where there is a treatment center recognized by the Cutaneous Lymphoma Foundation (CLF), and 64.87% originated in a city with a CLF-recognized treatment center. Conclusion: Valchlor remains a valuable therapy for CTCL; however, its rising cost and variable prescribing patterns warrant closer scrutiny, suggesting opportunities to improve prescribing practices and reduce overall costs. These data provide a foundation for further evaluation of real-world usage and can inform future value-based care strategies in CTCL management.

Peripheral T-cell lymphomas (PTCL) are aggressive hematologic cancers characterized by complex tumor microenvironments (TME) with rich hematopoietic and non-hematopoietic elements. A deeper understanding of both the malignant T cells and the TME could reveal therapeutic vulnerabilities in these lymphomas, which respond poorly to conventional chemotherapy. However, the cellular architecture, transcriptional programs and tumor cell-TME interactions have not been comprehensively characterized in PTCL. Here, we performed an integrated, multiomic analysis, including single-nucleus RNA sequencing (snRNA-seq), bulk RNA sequencing, and targeted mutational profiling, on archival samples from 29 frozen PTCL biopsies- including T follicular helper cell lymphoma (TFHL) and PTCL, not otherwise specified (PTCL, NOS), along with six control lymph nodes and tonsils. After rigorous quality control, a total of 230,174 cells were analyzed. Unsupervised clustering following dimensionality reduction revealed 13 distinct benign and malignant T-cell states and 42 transcriptionally unique cell types (14 B-cell subsets, 14 myeloid subsets, 6 endothelial cell subsets, and 8 non-endothelial stromal cell subsets), many of which have not been previously characterized in PTCL. Analysis of malignant T cells revealed substantial heterogeneity in PTCL, NOS, linked to chromosomal abnormalities, while TFHL tumors exhibited less genetic and transcriptional variability and were associated with specific transcriptional programs driven by RHOA G17V and IDH2 mutations. Both subtypes showed enrichment in MYC targets, TNF-α signaling, mTOR signaling, mitosis, and cell cycle pathways, whereas MAPK and STAT5 signaling were uniquely enriched in TFHL. TME analysis revealed shared hallmarks across PTCL, NOS and TFHL, including B-cell depletion and expansion of exhausted CD8+ T cells and tumor-associated macrophages (Mφ). In contrast, enrichment of immunoblasts, regulatory T cells, and non-endothelial stromal cells represents a disease-specific alteration in TFHL. Decomposition of transcriptomic profiles from all cell types present in PTCL and control samples using non-negative matrix factorization (NMF) identified 104 gene expression programs, 36 of which showed differential per-sample activity between PTCL TME and normal controls. Correlation analysis of PTCL-specific gene programs revealed five multicellular modules, including two non-overlapping modules: one characterized by follicular dendritic cells (FDCs) and inflammatory Mφ, and the other by cancer-associated fibroblasts (CAFs) and M2-like Mφ. Cell-cell communication analysis further identified TNF-α and TGF-β as key ligands orchestrating these TME modules. Finally, computational modeling and in vivo perturbation of these interactions uncovered therapeutic vulnerabilities, including a previously unrecognized role for the TGF-β pathway in supporting PTCL tumor growth. Together, these findings provide new insights into PTCL biology and establish a foundation for TME-informed therapeutic strategies, offering potential avenues for patient stratification and targeted intervention.

Gamma Secretase Inhibitors (GSIs) effectively block oncogenic Notch homolog-1 (NOTCH1), a characteristic feature of T cell acute lymphoblastic leukemias (T-ALL). However, their clinical application has been stalled by the induction of severe gastrointestinal toxicity resulting from the inhibition of NOTCH signaling in the gut, which translates into increased goblet cell differentiation. Genome-wide CRISPR loss-of-function screen in the colon cancer cell line LS174T identified the neddylation pathway as a main regulator of goblet cell differentiation upon NOTCH1 inhibition. Consistently, pharmacologic inhibition of the neddylation pathway with the small molecule inhibitor MLN4924, rescued GSI-induced differentiation in LS174T cells. Mechanistically, neddylation inhibition by MLN4924 increases the protein stability of Hairy and enhancer of split-1, a direct NOTCH1 transcriptional target and key regulator of absorptive and secretory cell fate decisions. Combined treatment with GSI and MLN4924 in a murine Notch1-dependent model of T-ALL led to leukemia regression and improved overall survival in the absence of gut toxicity. Overall, these results support the combined targeting of the NOTCH1 and neddylation pathways for the treatment of NOTCH1-induced T-ALL.

Transcriptional regulation, involving the complex interplay between regulatory sequences and proteins, directs all biological processes. Computational models of transcription lack generalizability to accurately extrapolate in unseen cell types and conditions. Here, we introduce GET, an interpretable foundation model designed to uncover regulatory grammars across 213 human fetal and adult cell types. Relying exclusively on chromatin accessibility data and sequence information, GET achieves experimental-level accuracy in predicting gene expression even in previously unseen cell types. GET showcases remarkable adaptability across new sequencing platforms and assays, enabling regulatory inference across a broad range of cell types and conditions, and uncovering universal and cell type specific transcription factor interaction networks. We evaluated its performance on prediction of regulatory activity, inference of regulatory elements and regulators, and identification of physical interactions between transcription factors. Specifically, we show GET outperforms current models in predicting lentivirus-based massive parallel reporter assay readout with reduced input data. In fetal erythroblasts, we identify distal (>1Mbp) regulatory regions that were missed by previous models. In B cells, we identified a lymphocyte-specific transcription factor-transcription factor interaction that explains the functional significance of a leukemia-risk predisposing germline mutation. In sum, we provide a generalizable and accurate model for transcription together with catalogs of gene regulation and transcription factor interactions, all with cell type specificity.

Cutaneous T-cell lymphomas are mature lymphoid neoplasias resulting from the malignant transformation of skin-resident T-cells. A distinctive clinical feature of cutaneous T-cell lymphomas is their sensitivity to treatment with histone deacetylase inhibitors. However, responses to histone deacetylase inhibitor therapy are universally transient and noncurative, highlighting the need for effective and durable drug combinations. In this study, we demonstrate that the combination of romidepsin, a selective class I histone deacetylase inhibitor, with afatinib, an EGFR family inhibitor, induces strongly synergistic antitumor effects in cutaneous T-cell lymphoma models in vitro and in vivo through abrogation of Jak-signal transducer and activator of transcription signaling. These results support a previously unrecognized potential role for histone deacetylase inhibitor plus afatinib combination in the treatment of cutaneous T-cell lymphomas.

Abstract Peripheral T-cell lymphomas (PTCL) are heterogeneous and highly aggressive hematologic malignancies with dismal outcomes, highlighting the need for novel targeted therapies. PTCLs are characterized by complex tumor-microenvironment (TME) ecosystems with low tumor cell content, admixed with rich immune infiltrates and stromal elements correlated with clinical outcomes. Direct targeting of the TME is an attractive strategy for PTCLs; however, the specific mechanisms that regulate the interdependence between lymphoma and TME remain poorly understood, hampering the development of TME-directed therapies. Our previous work based on driver genetic alterations in PTCL patients and genetic mouse models identified an important role of these driver mutations in remodeling the TME (Cortes et al., 2022). To address the specific role and mechanisms of tumor cell-TME interactions in the pathogenesis of PTCL, we developed a novel experimental platform and cutting-edge computational methodologies using single-nucleus RNAseq analysis to comprehensively profile lymphoma cells and their microenvironment from 27 PTCL patient samples of two major nodal pathologic subtypes, including 18 Nodal T follicular helper cell lymphomas (TFHcL) and 9 PTCL, not otherwise specified (PTCL, NOS). Our approach allowed the identification and in-depth characterization of tumor cells, lymphoid and myeloid TME, and vascular endothelium and non-endothelial stromal cells, which were challenging to capture with conventional single-cell RNAseq. The malignant T-cells in PTCL, NOS were grouped into patient-specific clusters demonstrating significant inter-tumor heterogeneity associated with diverse and frequently complex chromosomal abnormalities identified by copy number variation (CNV) analysis. In contrast, malignant TFHcL cells show a low degree to no apparent chromosomal abnormalities and divide into two major transcriptional subclusters, with tumor cells from the same patient falling into one of the two subclusters. Additionally, genomic profiles of TFHcL tumors revealed that the inter-tumor transcriptional heterogeneity of TFHcL tumor cells was strongly associated with the driver RHOA G17V mutation status. To profile oncogenic pathways critical for PTCL growth, we compared PTCL tumor cells to normal CD4 T-cells from reactive lymphoid tissue. PTCL, NOS and TFHcL tumor cells were equally enriched in the TCR signaling pathway, mitosis, and cell cycle related genes. However, exploitation of KRAS and cytokine-driven inflammatory signals was prevalent in TFHcL but not PTCL, NOS tumor cells. Analysis of the immune TME identified an increase in T regulatory cells, CD8 T-cells, immunoblasts, and macrophages in TFHcL. Furthermore, while dysregulated B-cells with heightened cytokine-driven inflammatory signals were unique to TFHcL, inflammatory macrophages and classic dendritic cells (cDC) were prominent in the TME of both subtypes. Our results uncover shared and subtype-specific pathogenic features in PTCL and pave the way for designing novel precision therapies for PTCL patients. Citation Format: Wen-Hsuan Wendy Lin, Anqi W Wang, Bobby B Shih, Jean-Baptiste F Reynier, Laura Quevedo Palacio, Craig Soderquist, Ryan D Najac, Cindy Ma, Govind Bhagat, Adolfo A Ferrando, Raul Rabadan, Teresa Palomero. Single-cell transcriptomics reveals shared and subtype-specific vulnerabilities of the tumor-microenvironment ecosystems in peripheral T-cell lymphomas [abstract]. In: Proceedings of the Fourth AACR International Meeting on Advances in Malignant Lymphoma: Maximizing the Basic-Translational Interface for Clinical Application; 2024 Jun 19-22; Philadelphia, PA. Philadelphia (PA): AACR; Blood Cancer Discov 2024;5(3_Suppl):Abstract nr PR05.

We describe the synthesis and biological testing of ruthenium-bipyridine ruxolitinib (RuBiRuxo), a photoreleasable form of ruxolitinib, a JAK inhibitor used as an antitumoral agent in cutaneous T-cell lymphomas (CTCL). This novel caged compound is synthesized efficiently, is stable in aqueous solution at room temperature, and is photoreleased rapidly by visible light. Irradiation of RuBiRuxo reduces cell proliferation and induces apoptosis in a light- and time-dependent manner in a CTCL cell line. This effect is specific and is mediated by a decreased phosphorylation of STAT proteins. Our results demonstrate the potential of ruthenium-based photocompounds and light-based therapeutic approaches for the potential treatment of cutaneous lymphomas and other pathologies.

Post-myeloproliferative neoplasm secondary Acute Myeloid Leukemia (Post-MPN sAML) is an aggressive and lethal hematologic malignancy arising from myeloproliferative neoplasms (MPN). Hyperactivation of JAK/STAT signaling is present in 50-90% of MPNs, underscoring its role in driving MPN pathogenesis and transformation to acute sAML. Ruxolitinib, a Type-I JAK2 inhibitor approved for the treatment of myelofibrosis, was tested in a Phase II clinical trial in post-MPN sAML. Despite enhancing patients' quality of life, ruxolitinib did not exert major improvements on disease outcome (Eghtedar et al, 2012). Thus, there is an urgent need to find synergistic drug combinations capable of complementing the therapeutic activity of JAK2 inhibitors in post-MPN sAML. Here, we demonstrate that combination treatment of ruxolitinib and CBP30, a bromodomain inhibitor of histone acetyltransferases (HAT) CREBBP and p300, enhances the therapeutic activity of ruxolitinib in post-MPN sAML using high throughput genetic screens, in vitro functional assays, transcriptomic and epigenetic profiling and mouse xenograft in vivo experiments. To identify druggable determinants of JAK2 inhibitor response in post-MPN sAML, we performed genome-wide loss-of-function CRISPR screens in sAML HEL cells treated with four different JAK2 inhibitors, including ruxolitinib. We discovered that depletion of the histone acetyl transferase CREBBP sensitizes HEL cells to JAK2 inhibition (Figure 1A). Pharmacological inhibition of JAK/STAT signaling and HAT activity using a combination of ruxolitinib and CREBBP inhibitors significantly reduced the proliferation of sAML cells HEL and SET2 (Figure 1B). Among the CREBBP inhibitors, CBP30 synergizes best with ruxolitinib having Chou-Talalay combination indices of 0.152 and 0.704 in HEL and SET2, respectively. Moreover, combination treatment of ruxolitinib and CBP30 markedly increased apoptosis and induced cell cycle arrest at G1 in HEL. To mechanistically characterize the synergism between ruxolitinib and CBP30, we performed ChIP-seq to profile STAT5 binding and H3K27ac levels in HEL cells treated with DMSO, ruxolitinib, CBP30 and their combination. STAT5 binding profiles did not reveal major differences in ruxolitinib versus ruxolitinib + CBP30 treated cells, suggesting that the drug combination does not significantly alter STAT5 binding. Conversely, we noticed striking global changes in H3K27ac levels in the combination compared with DMSO or single agents, indicating that the synergistic effect between ruxolitinib and CBP30 is mediated by substantial changes in transcriptional regulation. Notably, analysis of H3K27ac downregulated and upregulated sites identified enrichments in motifs recognized by FOS/SMAD4 and GATA transcription factors, respectively. These TFs are involved in the regulation of myeloid differentiation implicating a synergistic regulation of relevant myeloid differentiation pathways by the combination of JAK2 and CREBBP inhibitors in sAML. Furthermore, ruxolitinib + CBP30 treatment in HEL elicits a unique transcriptome profile with expression deregulation of genes associated with cell cycle, DNA repair, HAT activity and TGF-beta signaling as determined by DAVID functional annotation analysis. Lastly, to evaluate the synergistic effects of ruxolitinib and CBP30 in vivo, we treated luciferase-expressing HEL sAML mouse xenografts with vehicle, ruxolitinib (80 mg/kg), CBP30 (30 mg/kg) and ruxolitinib (80 mg/kg) + CBP30 (30 mg/kg). Combination of ruxolitinib and CBP30 displayed significant decrease in leukemia burden after 4 weeks of treatment as monitored by bioluminescence imaging and quantification. Consistently, mice treated with both drugs also showed remarkable decrease in their white blood cell count and spleen size compared with single-agent-treated and vehicle-treated mice. Collectively, our results substantiate the combinatorial therapeutic targeting of JAK/STAT signaling and HAT activity as a potential treatment approach for post-MPN sAML.

Post-MyeloProliferative Neoplasm Acute Myeloid Leukemia (Post-MPN AML) is an aggressive and lethal hematologic malignancy arising from myeloproliferative neoplasms. Currently, there is no standard of care treatment for post-MPN AML patients, with allogeneic stem cell transplantation being the only curative option. However, the advanced age and comorbidities of post-MPN AML patients render them unfit for allo-SCT, emphasizing the urgent unmet need to find novel treatment for this disease. Hyperactivation of JAK/STAT signaling is highly prevalent in post-MPN AML pathogenesis, with as much as 50-90% of MPN cases harboring driver mutations in JAK2, CALR or MPL. The JAK2 inhibitor ruxolitinib has been tested in clinical trials for post-MPN AML and was reported to improve the patient quality of life, however, it was not effective in modifying the course of the disease. Given the limitations of targeting the JAK/STAT pathway, we aimed to identify genes and pathways synergistic with JAK2 inhibitors in post-MPN AML. Here, we demonstrate through genome-wide CRISPR screens in post-MPN AML line HEL treated with four different JAK2 inhibitors (i.e. ruxolitinib, momelotinib, pacritinib and fedratinib) that depletion of CREBBP sensitizes cells to JAK2 inhibition. CREBBP is a histone acetyltransferase which also acts as a coactivator of essential transcription factors in diverse hematopoietic lineages. CREBBP is overexpressed in AML and is ubiquitously expressed in blood cancer cell lines, underscoring its potential as a therapeutic target. Genetic depletion of CREBBP via CRISPR/Cas9 editing showed that CREBBP, but not its paralog EP300, sensitizes HEL to JAK2 inhibition. In addition, pharmacological approaches showed that JAK2 inhibitors synergize with CREBBP/EP300 inhibitors to effectively kill post-MPN AML lines. In both human and murine models of post-MPN AML, the combination treatment of ruxolitinib plus CREBBP/EP300 inhibitor SGC-CBP30 or CCS1477 substantially induced apoptosis and cell cycle arrest at G1. Moreover, bioluminescence imaging of HEL luciferase mouse xenografts treated with vehicle, ruxolitinib (80 mg/kg), CCS1477 (20mg/kg) and combination showed significant decrease in leukemia burden in the combination-treated animals after 3 weeks of treatment. Integration of the CRISPR viability screen and transcriptome profiles of HEL cells treated with JAK and CREBBP inhibitors identified 118 genes that are both essential for the proliferation and are downregulated with the combination treatments. GSEA showed enrichment of MYC targets in the said genes with E2F4, MYC and HSF1 being the top transcription factors associated with them. Notably, 21 of the 118 genes identified are part of the epichaperome network which is regulated by MYC. Accordingly, western blot analysis showed decreased MYC, E2Fs, STAT3 and STAT5 protein levels in HEL cells treated with the combinations. In addition, combination treatments also reversed the accumulation of JAK2 protein levels induced by ruxolitinib, which is a well-studied mechanism driving disease persistence. We hypothesize that this is due to the downregulation of MYC-regulated epichaperome genes which are known to regulate JAK/STAT signaling by physically interacting and stabilizing JAK2. Further investigation on the epigenetic changes induced by the ruxolitinib/CREBBP inhibitors is underway to identify their modulatory role of MYC and STAT3/5 regulatory regions. Overall, our results demonstrate that CREBBP/EP300 inhibition potentiates JAK2 inhibition in post-MPN AML by further attenuating MYC expression and activity, and repressing JAK/STAT and other pathways associated with JAK2 inhibitor persistence. Therefore, we propose CREBBP inhibition as a potential therapeutic strategy to potentiate JAK2 inhibition in post-MPN AML.

T-cell Acute Lymphoblastic Leukemia (T-ALL) is hematologic tumor characterized by the diffuse infiltration of the bone marrow by malignant hematopoietic cells expressing immature T cell markers. Although T-ALL currently has better cure rates primarily due to multiagent or intensified chemotherapy, the prognosis for patients who are resistant or develop relapse to therapy remains very poor. Aberrant Notch homolog-1 (NOTCH1) signaling is a major driver of T-ALL pathogenesis as more than 60% of T-ALL cases harbor activating mutations in the NOTCH1 gene. Gamma Secretase Inhibitors (GSIs) which effectively block the activation of oncogenic protein NOTCH1 are potential candidates for the treatment of T-ALL. However, the clinical application of GSIs is hampered by severe gastrointestinal toxicity due to the inhibition of NOTCH1 signaling in the gut. Here we demonstrate that combination therapy of GSIs and a small molecule inhibitor of the neddylation pathway circumvents the GSI-induced gut toxicity in vitro and in vivo. Genome-wide CRISPR loss-of-function screen in LS174T adenocarcinoma cells revealed neddylation pathway as a main regulator of massive goblet cell differentiation upon NOTCH1 inhibition. Genetic and pharmacologic inhibition of the neddylation pathway in LS174T cells rescued GSI-induced differentiation and cell death. Mechanistically, neddylation inhibition increases the protein stability of Hairy and enhancer of split-1 (HES1), a known regulator of absorptive and secretory cell fate decisions. Combination treatment of GSI and neddylation inhibitor in C57/BL6 mice showed a profound decrease in the number of goblet cells and maintained HES1 protein levels in the intestinal epithelium compared to GSI treatment alone. Remarkably, combined treatment of GSI and neddylation inhibitor in NOTCH1-induced T-ALL mice showed leukemia regression and improved overall survival without any associated gut toxicity. Overall, these results substantiate the potential of targeting NOTCH1 and neddylation pathway in the treatment of NOTCH1-induced T-ALL. Citation Format: Carla Bertulfo, Pablo Perez-Duran, Teresa Palomero, Adolfo Ferrando. Therapeutic targeting of NOTCH1 and neddylation pathway in T cell acute lymphoblastic leukemia. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3928.

Peripheral T cell lymphomas (PTCLs) are a group of aggressive lymphoid malignancies originating from mature T cells. Angioimmunoblastic T-cell lymphomas (AITL) represent 20% to 30% of all PTCL diagnoses and are associated with autoimmune features, poor response to chemotherapy, and dismal prognosis. Molecular profiling of AITL has identified T-follicular helper (TFH) cells as the cell of origin of AITL, while mutational analysis has identified frequent alterations in epigenetic regulators ( TET2, DNMT3A, and IDH2) and elements of the TCR pathway as drivers of AITL transformation. Our group identified the highly recurrent RHOA G17V mutation as a defining hallmark of AITL and other T-cell lymphomas of TFH-cell origin (Palomero et al., 2014). Using a novel conditional knockin mouse model, we demonstrated that expression of Rhoa G17V in CD4+ T-cells induces TFH cell specification and promotes AITL lymphomagenesis in the context of loss of Tet2 (Cortes et al., 2018). Recently, we have identified that Rhoa G17V regulates the expression of the sphingosine-1-phosphate (S1P) receptor 1 (S1PR1) one of the five members of the receptor family of S1P, which plays an essential role in immune response and lymphocyte trafficking. During thymic T-cell development, expression of Rhoa G17V in CD4+ cells inhibited S1PR1 downregulation after the double positive stage leading to altered positive selection, reduced CD4+ single positive cell numbers and premature thymic egress of autorreactive T-cells which results in the development of systemic inflammation and autoimmunity. In our Tet2 −/− RHOA G17V AITL tumor model, constitutive activation of S1P receptors supported activation of the STAT3 and NFkB pathway and enhanced migration of malignant cells to peripheral organs. Mechanistically, expression of Rhoa G17V in mature CD4+ T-cells impaired re-phosphorylation of the Ezrin-Radoxin-Moesin (ERM) complex independently of ROCK signaling and led to the activation and increased migration of CD4+ T-cells. Importantly, blockade of S1PR1 signaling by fingolimod induced anti-tumor activity in Tet2 −/− RHOA G17V lymphoma-bearing mice in vivo with significant reductions in tumor burden associated with increased apoptosis and decreased STAT3 phosphorylation. Furthermore, analysis of a panel of primary tumors from AITL patients using a multiplexed staining panel and PhenoImager platform revealed significantly higher S1PR1 expression in AITL tumor cells than normal TFH cells from benign reactive lymph nodes. Our findings highlight the role of S1P signaling pathway in the pathogenesis of RHOA G17V-driven AITLs and the potential therapeutic benefit of targeting this pathway.

The human bone marrow (BM) is one of the most complex and critical tissues in the adult, functioning as the site for hematopoietic stem and progenitor cell (HSPC) maintenance, as well as blood and immune cell production in homeostasis, injury, and disease. As the development of human organs-on-a-chip (OoC) platforms has emerged over the past decade, there has been an increased relevance of using human BM models to study human-specific immune interactions in vitro. Here we report the development of a patient-specific bioengineered model of the BM, derived entirely from induced pluripotent stem cells (iPSCs) and its use in studies of radiation toxicity, cancer, and systemic injury responses. The engineered model of human BM (eBM) is derived from iPSC-derived osteoblasts, mesenchymal stem/stromal cells, and endothelial cells within a decellularized bone scaffold. This model was developed to be modular, with ability to include either healthy cord blood-, iPSC-, or BM-derived HSPCs. We validated the model with histological staining, flow cytometry, colony forming assays, and single-cell RNA sequencing, demonstrating maintenance and differentiation of blood progenitors and progeny, and describe here applications in acute radiation injury, leukemic infiltration, solid tumor metastasis, and systemic, multi-tissue interactions. Notably, eBMs were able to maintain donor acute myeloid and B-/T-lymphoblastic leukemias better than in liquid cultures, as well as recapitulating drug responses seen clinically. Further, we linked the healthy eBM model to engineered cardiac tissues by vascular perfusion to demonstrate the recruitment of monocytes from eBMs in response to an acute cardiac injury, in a multi-OoC setting. We propose that this novel model system can be used to study malignant transformation, systemic immune responses, and the development of personalized therapeutics. The human bone marrow (BM) is one of the most complex and critical tissues in the adult, functioning as the site for hematopoietic stem and progenitor cell (HSPC) maintenance, as well as blood and immune cell production in homeostasis, injury, and disease. As the development of human organs-on-a-chip (OoC) platforms has emerged over the past decade, there has been an increased relevance of using human BM models to study human-specific immune interactions in vitro. Here we report the development of a patient-specific bioengineered model of the BM, derived entirely from induced pluripotent stem cells (iPSCs) and its use in studies of radiation toxicity, cancer, and systemic injury responses. The engineered model of human BM (eBM) is derived from iPSC-derived osteoblasts, mesenchymal stem/stromal cells, and endothelial cells within a decellularized bone scaffold. This model was developed to be modular, with ability to include either healthy cord blood-, iPSC-, or BM-derived HSPCs. We validated the model with histological staining, flow cytometry, colony forming assays, and single-cell RNA sequencing, demonstrating maintenance and differentiation of blood progenitors and progeny, and describe here applications in acute radiation injury, leukemic infiltration, solid tumor metastasis, and systemic, multi-tissue interactions. Notably, eBMs were able to maintain donor acute myeloid and B-/T-lymphoblastic leukemias better than in liquid cultures, as well as recapitulating drug responses seen clinically. Further, we linked the healthy eBM model to engineered cardiac tissues by vascular perfusion to demonstrate the recruitment of monocytes from eBMs in response to an acute cardiac injury, in a multi-OoC setting. We propose that this novel model system can be used to study malignant transformation, systemic immune responses, and the development of personalized therapeutics.

Supplementary Table from Pharmacologic Inhibition of NT5C2 Reverses Genetic and Nongenetic Drivers of 6-MP Resistance in Acute Lymphoblastic Leukemia

Peripheral T cell lymphoma not otherwise specified (PTCL-NOS) comprises heterogeneous lymphoid malignancies characterized by pleomorphic lymphocytes and variable inflammatory cell-rich tumor microenvironment. Genetic drivers in PTCL-NOS include genomic alterations affecting the VAV1 oncogene; however, their specific role and mechanisms in PTCL-NOS remain incompletely understood. Here we show that expression of Vav1-Myo1f, a recurrent PTCL-associated VAV1 fusion, induces oncogenic transformation of CD4+ T cells. Notably, mouse Vav1-Myo1f lymphomas show T helper type 2 features analogous to high-risk GATA3+ human PTCL. Single-cell transcriptome analysis reveals that Vav1-Myo1f alters T cell differentiation and leads to accumulation of tumor-associated macrophages (TAMs) in the tumor microenvironment, a feature linked with aggressiveness in human PTCL. Importantly, therapeutic targeting of TAMs induces strong anti-lymphoma effects, highlighting the lymphoma cells' dependency on the microenvironment. These results demonstrate an oncogenic role for Vav1-Myo1f in the pathogenesis of PTCL, involving deregulation in T cell polarization, and identify the lymphoma-associated macrophage-tumor microenvironment as a therapeutic target in PTCL.

Abstract Low-intensity maintenance therapy with 6-mercaptopurine (6-MP) limits the occurrence of acute lymphoblastic leukemia (ALL) relapse and is central to the success of multiagent chemotherapy protocols. Activating mutations in the 5′-nucleotidase cytosolic II (NT5C2) gene drive resistance to 6-MP in over 35% of early relapse ALL cases. Here we identify CRCD2 as a first-in-class small-molecule NT5C2 nucleotidase inhibitor broadly active against leukemias bearing highly prevalent relapse-associated mutant forms of NT5C2 in vitro and in vivo. Importantly, CRCD2 treatment also enhanced the cytotoxic activity of 6-MP in NT5C2 wild-type leukemias, leading to the identification of NT5C2 Ser502 phosphorylation as a novel NT5C2-mediated mechanism of 6-MP resistance in this disease. These results uncover an unanticipated role of nongenetic NT5C2 activation as a driver of 6-MP resistance in ALL and demonstrate the potential of NT5C2 inhibitor therapy for enhancing the efficacy of thiopurine maintenance therapy and overcoming resistance at relapse. Significance: Relapse-associated NT5C2 mutations directly contribute to relapse in ALL by driving resistance to chemotherapy with 6-MP. Pharmacologic inhibition of NT5C2 with CRCD2, a first-in-class nucleotidase inhibitor, enhances the cytotoxic effects of 6-MP and effectively reverses thiopurine resistance mediated by genetic and nongenetic mechanisms of NT5C2 activation in ALL. This article is highlighted in the In This Issue feature, p. 2483

Aging is characterized by an accumulation of myeloid-biased hematopoietic stem cells (HSCs) with reduced developmental potential. Genotoxic stress and epigenetic alterations have been proposed to mediate age-related HSC loss of regenerative and self-renewal potential. However, the mechanisms underlying these changes remain largely unknown. Genetic inactivation of the plant homeodomain 6 (Phf6) gene, a nucleolar and chromatin-associated factor, antagonizes age-associated HSC decline. Immunophenotyping, single-cell transcriptomic analyses and transplantation assays demonstrated markedly decreased accumulation of immunophenotypically defined HSCs, reduced myeloid bias and increased hematopoietic reconstitution capacity with preservation of lymphoid differentiation potential in Phf6-knockout HSCs from old mice. Moreover, deletion of Phf6 in aged mice rejuvenated immunophenotypic, transcriptional and functional hallmarks of aged HSCs. Long-term HSCs from old Phf6-knockout mice showed epigenetic rewiring and transcriptional programs consistent with decreased genotoxic stress-induced HSC aging. These results identify Phf6 as an important epigenetic regulator of HSC aging.

Acute leukemias represent the most frequent type of cancer (∼30%) in children and young adults. The lack of robust systems for in vitro culture of primary leukemia samples is a significant barrier for the development of effective genetic and chemical screens for novel therapeutic targets in pediatric leukemia. In vitro systems, including engineered tissues and organ-on-a-chip systems, are of increasing interest in the stem cell and cancer fields. In vitro models of the bone marrow (BM) have yet to gain momentum, largely due to their reduced throughput, technical complexity, and heterogeneity of starting stromal cell populations. In this work, we present a bioengineered, human, induced pluripotent stem cell (iPSC)-derived bone marrow tissue model (eBM), comprised of osteoblasts, mesenchymal stromal cells, and endothelial cells within a decellularized bone scaffold. This model was developed to include both healthy BM-derived hematopoietic stem and progenitor cells (HSPCs) and acute lymphoblastic leukemia (ALL) blasts. We first demonstrate the ability of our eBM tissue model to support ALL blast maintenance in vitro over 4 weeks, with survival of patient-derived xenograft samples significantly higher (5-10 fold expansion) than in monolayer controls. Acute leukemias are known to alter their microenvironmental niche, and in many cases, use the stroma to protect malignant clones during treatment. In this modular eBM model, the combination of all supporting populations (osteoblasts, endothelium, and mesenchymal stromal cells) was crucial in supporting difficult-to-culture ALL donor cells, likely because of the continuous, responsive secretion of hematopoietic regulatory factors (i.e. osteopontin, IL-7, CXCL8). We propose that this novel model system will advance studies of the human BM during malignant transformation and in the development of personalized therapeutics.