Yang, Zhi-Zhang

Abstract The role of the bone marrow (BM) microenvironment in regulating the antitumor immune response in Waldenstrom macroglobulinemia (WM) remains limited. Here we transcriptionally and phenotypically profiled non-malignant (CD19− CD138−) BM cells from WM patients with a focus on myeloid derived suppressive cells (MDSCs) to provide a deeper understanding of their role in WM. We found that HLA-DRlowCD11b+CD33+ MDSCs were significantly increased in WM patients as compared to normal controls, with an expansion of predominantly polymorphonuclear (PMN)-MDSCs. Single-cell immunogenomic profiling of WM MDSCs identified an immune-suppressive gene signature with upregulated inflammatory pathways associated with interferon and TNF signaling. Gene signatures associated with an inflammatory and immune suppressive environment were predominately expressed in PMN-MDSCs. In vitro WM PMN-MDSCs demonstrated robust T-cell suppression, and their viability and expansion was notably enhanced by G-CSF and TNFα. Furthermore, BM malignant B-cells attracted PMN-MDSCs to a greater degree than monocytic MDSCs. Collectively, these data suggest that malignant WM B cells actively recruit PMN-MDSCs which promote an immunosuppressive BM microenvironment through a direct T cell inhibition, while release of G-CSF/TNFα in the microenvironment further promotes PMN-MDSC expansion and in turn immune suppression. Targeting PMN-MDSCs may therefore represent a potential therapeutic strategy in patients with WM.

Despite notable advancements in treating chronic lymphocytic leukemia (CLL) with targeted therapies, a significant clinical challenge arises as a subset of CLL patients progress and develop Richter's transformation (RT), an aggressive form of lymphoma. While immune cell dysfunction and T cell defects are well-known characteristics of CLL, the alterations in immune cell subsets during CLL progression to RT remain largely unexplored. While CLL patients typically fail to respond to anti-PD1 pembrolizumab, a positive response has been observed in 20-40% of RT patients. Previously, our study involving three RT patients (1 responder and 2 progressors) revealed enriched regulatory T cells (Tregs) in the responders compared to progressors after pembrolizumab treatment (Iyer et al, ASH2020). Hence, we propose that dynamic changes in T cell subsets at baseline may be associated with response to pembrolizumab treatment. To this end, we performed single-cell RNA sequencing on sorted T cells (CD3 +) from PBMCs derived from 20 patients (responders (RES), n=5; stable disease (SD), n=5; progressive disease/progressors (PD), n=10). A total of 408,627 cells were captured, and an average of 1,393 genes were detected per cell per patient. Using a graph-based clustering approach (Seurat V4.3), we identified 19 clusters based on marker genes and transcriptional signatures and identified enriched pathways using Enrichr analysis. Our results showed that RES had an increased Th17 subset at baseline compared to PD (15% vs. 5%, p=0.01) and reduced Th2 T cells compared to SD (p=0.004) and PD (p=0.05). The Th17/Th2 ratio was higher in RES at baseline (Th17/Th2 R: 1.24, PD: 0.21, SD: 0.25, p0.05) and post-pembrolizumab (Th17/Th2 RES: 1.31, PD: 0.34, SD: 0.30, p0.05). Furthermore, RES had an increased fraction of proliferating CD4 T cells at baseline (4% RES vs. 0.4% SD and 0.4% PD, p=0.02) and post-treatment (1.9% RES vs. 0.8% PD and 0.34% SD, p=0.02). Although we observed an increasing trend in the fraction of Tregs in the RES at baseline [(7.6% RES vs.6.1% PD and 6.5% SD, p=0.4; ns) and post-treatment (10.3% RES vs. 8.3% PD and 5.8% SD, p= 0.2; ns] compared to PD and SD, the difference was not significant. Similarly, among CD8+ T cells, we observed an increasing trend in CD8+ effector T cell fraction in the responders at baseline (29.8% RES vs. 21% PD and 15.6% SD, p=0.9, ns) and post-treatment (30.1% RES vs. 21.2% PD and 19.9% SD, p=0.44, ns). However, the difference was not statistically significant. These findings highlight the dynamic changes in T cell subsets associated with pembrolizumab response. In terms of transcriptional changes in Tregs following pembrolizumab treatment among RES, PD, and SD, we observed dynamic changes on a few Treg marker genes, including TIGIT and TOX. TIGIT and TOX expression drastically decreased after treatment in RES (p < 0.0001), while expression did not change in PD and SD. There were no significant changes in the expression of FOXP3 or CD27 in RES, PD, and SD before and after the treatment. In CD8 effector T cells, GZMK, GZMA (granzymes), and PRF1 (perforin) are marker genes. We observed higher expression of GZMK and GZMA in RES (p <0.0001) at the baseline compared to SD and PD. In the proliferating CD4 T cells, we observed increased expression of IDH2 (Isocitrate dehydrogenase 2-key enzyme of TCA cycle) in RES at the baseline, as compared to PD and SD, indicating increased TCA cycle activity in the proliferating T cells. Pathway analysis of T cell clusters using the Enrichr database showed TNFA-IFNG-MYC targets are enriched in CD8 + cells, IL2-STAT5 signaling is upregulated in Th17, TREGs, and E2F cell cycle associated pathways enriched in CD4 + proliferation cluster. In conclusion, our study unveils significant baseline differences in CD4 subsets between RES and PD/SD subjects and dynamic changes in T cell subsets and transcriptional signatures. These findings offer potential predictors of pembrolizumab response in RT, such as increased Th17/Th2 ratio, enriched proliferating CD4 + T cells, and elevated TIGIT, TOX gene expression in Tregs, GZMK, and GZMA in CD8 effector cells. This deeper understanding of the immunological landscape in RT patients sheds light on the interplay between T cell subsets and treatment response in RT, paving the way for personalized therapeutic strategies. Further validation and exploration can lead to improved outcomes and tailored interventions for RT.

Introduction: Waldenström macroglobulinemia (WM) is a rare B-cell lymphoma that may be preceded by an asymptomatic IgM-monoclonal gammopathy of undetermined significance (MGUS). The mechanisms underlying the progression from IgM-MGUS to WM remain to be fully elucidated; however, the stepwise accumulation of genomic abnormalities has been suggested as a potential driver. Currently, the role of epigenetic regulation of DNA remains underexplored in this disease spectrum. Aberrant DNA methylation is a hallmark of oncogenesis that plays an important role in hematologic malignancies. The aim of this study was to assess differential DNA methylation in WM compared to IgM-MGUS and identify hyper- and hypomethylated pathways that may alter key cellular functions. Methods: This study included 15 patients with WM and 6 with IgM-MGUS. Bone marrow samples were prospectively collected and CD19+ and/or CD138+ B cells were positively selected. Subsequently, genome-wide DNA methylation analysis was performed. CpG methylation ratios were segmented into 200-bp regions and differentially methylated regions (DMRs) comparing WM to IgM-MGUS were identified. DMRs with a Q-value <0.05 and absolute methylation difference of ≥20% were considered significant and utilized for analysis. DMRs were annotated according to genomic regions of interest (promotors, 3'UTR, 5'UTR, introns, exons, CpG islands, CpG shores). Next, genes with significant promotor hyper- or hypomethylation were assessed using the Ingenuity Pathway Analysis. Pathways with a corrected p-value<0.05 were considered statistically significant and a Z-value was calculated based on the degree of differential methylation. Results: Genome-wide DNA methylation analysis revealed a significant number of DMRs (n= 15691, hyper: 8248; hypo: 7443) in WM compared to IgM-MGUS. Assessing genomic regions, promotors (hyper: 1379; hypo: 2403), introns (hyper: 1548; hypo: 2292), CpG shores (hyper: 587; hypo: 849) and 3'UTR (hyper: 66; hypo: 193) regions were observed to have a significantly greater number of hypomethylated DMRs (all comparisons, p<0.001) in WM compared to IgM-MGUS (Figure A). While exons (hyper: 1686; hypo: 1378), CpG islands (hyper: 2128; hypo: 82), and 5'UTR (hyper: 854; hypo: 246) regions had a significantly greater number of hypermethylated DMRs (all comparisons, p<0.001) (Figure A). Next, promotor methylation-based pathway analysis revealed differential methylation of multiple cellular pathways important in WM pathogenesis (Figure B). These included pathways involved in intracellular signaling and the regulation of cellular proliferation, such as the STAT3 and CXCR4 pathways which were hypomethylated, and the PTEN signaling pathway which was hypermethylated. Of interest, we have previously demonstrated the transcription factor, STAT3, to have an increased expression in WM and to be implicated in IgM secretion (Elsawa, 2011). Next, multiple cytokine signaling pathways were observed to be hypomethylated in WM compared to IgM-MGUS including members of the interleukin (IL) family (IL-1, IL-8, IL-15, IL-23), VEGF, and TGF-β. Of relevance, previous studies have demonstrated the elevation of IL-8, IL-15, and increased expression of angiogenic cytokines, including VEGF, in patients with WM (Lewicki, 2010; Anagnostopoulos, 2007). Furthermore, our group has also shown miRNA-based epigenetic regulation of IL-8 and IL-15 in WM compared to IgM-MGUS (Chohan, 2023). Lastly, several pathways critical in the regulation of the cell structure and the extracellular matrix (ECM) were found to be hypomethylated including the FGF, actin cytoskeleton, PAK, and FAK signaling pathways. Of note, the FGF/FGFR axis has been demonstrated to be dysregulated in WM, and its targeting has been shown to halt WM cell growth (Sacco, 2021). Conclusions: This study demonstrates differential DNA methylation of key genomic regions in WM compared to IgM-MGUS. Promotor methylation-based analysis reveals hypomethylation of several important cellular pathways involved in cytokine signaling, cell structure/ECM, and intracellular signaling. These results underscore the influence of epigenetics in the IgM-gammopathy disease spectrum and suggest the potential role of aberrant DNA methylation underlying the malignant progression from IgM-MGUS to WM.

Reprogramming of cellular metabolism is a hallmark of cancer leading to improved cellular fitness and providing a selective advantage to malignant cells. Recently, we found that follicular lymphoma (FL) patients with poor prognosis display increased transcriptional and protein expression of IRF4 (IRF4 high), altered immune signaling and an immunosuppressive tumor microenvironment (TME). Furthermore, high expression of IRF4 rewired B cell metabolism towards glycolysis ( Mondello P et al, Blood 2022;140 Suppl 1:168-169). However, it's unknown whether this metabolic reprogramming may be responsible in part for the impairment of immune cells in the TME, which also rely on glucose for their proliferation and function. Using RNA-seq on 88 FL patients, we found a negative correlation between IRF4-regulated genes associated with glycolysis and tumor infiltration by major immune cell subsets in IRF4 high compared to IRF4 low tumors ( Fig 1A), suggesting that IRF4-driven tumor glycolysis may reinforce an immune suppressed phenotype. Especially, we observed that ALDOA and GPI genes were inversely correlated with expression of CD4 + T cells (r=0.34, p=0.003; and r=0.38, p=0.015, respectively) and T reg cells (r=0.3, p=0.05; and r=0.35, p=0.027, respectively) in IRF4 high but not in IRF4 low tumors. As radioactive glucose is the cornerstone of PET assessment, we investigated the correlation between baseline SUV values and IRF4 expression. Remarkably, we found significantly higher levels of FDG uptake in FL IRF4 high compared to IRF4 low cases (p= 0.003), supporting the glycolytic features of Iymphoma expressing IRF4. To define the metabolic activity and function of lymphoma cells with different IRF4 status, we performed seahorse assay in TMD8 and SUDHL2 cells (lymphoma cells with IRF4 high expression) transfected with siIRF4 or siCtr for 72 hours. In line with the IRF4-dependence on glycolysis, TMD8 transfected with siCtr had a low baseline glycolytic ECAR, while the same value doubled after silencing IRF4. Notably, after adding oligomycin, which stimulates glycolysis, TMD8 control cells showed a lower glycolytic capability compared with those deprived of IRF4, probably due to the already maximal addiction to glycolysis ( Fig 1B left). We then investigated the glycolytic demand of germinal center (GC) B cells differently expressing IRF4. Using irf4 fl/fl mice ( Klein et al, Nat Immunol 2006) crossed with Cγ 1cre animals ( Cγ 1cre; irf4 fl/fl) to induce conditional deletion of irf4 in GC B cells, we tested ex-vivo sorted GC B cells from Cγ 1cre; irf4 fl/fl, Cγ 1cre; irf4 fl/+ and WT mice immunized for 10 days. This time point is critical for studying IRF4 as it corresponds to the GC reaction peak when GC B cells express elevated IRF4 levels promoting terminal differentiation to plasma cells. As expected, GC B cells minimally relied on glycolysis and showed a low ability to respond to oligomycin in WT mice ( Weiser et al, Nat Immunol 2020). However, GC B cells with partial or complete loss of irf4 demonstrated a superior ability to response to increased glycolytic demand ( Fig 1B right), further supporting the conclusion that IRF4 influences B cell dependency on glycolysis. As the malignant cell metabolism can alter nutrient availability in the TME, we explored if lymphoma cells differently expressing IRF4 may change glucose levels in the media and in turn impact T cells activity. We found low glucose and high lactate levels in the media of cell lines with IRF4 high expression, while the opposite was observed for those that were IRF4 low. We then cocultured primary human CD4 + T cells with TMD8 cells transfected with either siIRF4 or siCtr for 3 days at three different concentrations of D-glucose: high (22.2 mM, Gluc hi), normal (5.5 mM, Gluc nor) or absent (0 mM, Gluc nil). In line with our prior findings, coculture of B cells in which IRF4 was knocked out showed a significant decrease in abundance of T reg cells and an increase in T FH cells compared to control in presence of Gluc hi or Gluc nor. In contrast, there was no difference in the overall amount of either T cell subtype irrespective of IRF4 status with Gluc nil, supporting the prerequisite of microenvironmental glucose as primary fuel for T cell survival and function. In conclusion, IRF4 rewires the metabolism in B cells with profound functional implications on the TME. Our data suggest that glucose metabolism represents a functional vulnerability that might be exploited therapeutically.

A substantial subset of follicular lymphoma (FL) patients has an early relapse with a poor outcome. We showed that this unfavorable group displays increased expression of IRF4, dysregulated immune signaling and an immunosuppressive microenvironment. At the molecular level, IRF4 controls expression of antigen presentation and immune molecules (e.g. HLA-DR, CD40, PDL1). Furthermore, silencing of IRF4 triggers anti-lymphoma immunity ( Mondello P et al, Blood 2022;140 Suppl 1:168-169). However, little is known on the molecular mechanisms whereby IRF4 controls the crosstalk between B and T cells. To address this question, first we performed RNA-seq, ATAC-seq and CUT&TAG of active (H3K27Ac, H3K4me3) and repressive (H3K27me3, H3K4me1) histone marks in TMD8 and HBL1 (lymphoma cells with high IRF4 expression) transfected with two pairs of siIRF4 or siCtr for 72 hours. Gene expression revealed a broad change in transcription, with 432 upregulated genes and 136 downregulated genes in siIRF4 compared to siCtr. Notably, the genes with increased expression following IRF4 silencing coded for antigen presentation, immune molecules and cytokines (e.g. HLA-DR, CD86, IL6, IL10). Accordingly, we observed significant enrichment of antigen presentation, interferon and immune response signatures ( Fig 1A). Integrating RNA-seq with chromatin accessibility showed a significant shift towards transcriptional upregulation at genes nearest to upregulated ATAC peaks, with significant enrichment at enhancers (n=771). Conversely, transcriptional suppression was enriched at enhancers (n=2,458) of downregulated peaks. k-means analyses of all union peaks, generated by merging ATAC with CUT&TAG peaks, revealed a significant gain of H3K4me3, with concordant loss of H3K4me1 at enhancers with upregulated gene expression. In contrast, H3K4me3 was severely decreased at suppressed peaks. TF motif analysis revealed an enrichment for PU.1 and E2A at IRF4-bound regions, suggesting their cooperation in regulating immune signaling. Since antigen presentation and other immune molecules are critical components of the immune synapse, we explored the role of IRF4 on the interaction between germinal center (GC) B and T cells by crossing irf4 fl/fl mice with conditional deletion of irf4 ( Klein et al, Nat Immunol 2006) and Cγ 1cre mice ( Cγ 1cre; irf4 fl/fl) to induce recombination in GC B cells. After 10 days of immunization (GC reaction peak), Cγ 1cre; irf4 fl/fl and Cγ 1cre; irf4 fl/+ GC B cells showed significant increase of HLA-DR and CD40 expression compared to WT. As expected, Cγ 1cre; irf4 fl/fl mice displayed a significantly skewed light-zone:dark-zone ratio in favor of increased dark-zone, which are the cells that most reflect the action of BCL6, with concordant suppression of plasma cells. Additionally, we observed a significant increase of CD4 + T cells after partial or complete deletion of irf4, with upregulation of T FH cells and decrease of T reg cells ( Fig 1B). Notably, mice with loss of irf4 showed a significant decrease in exhausted T FH cells. To test if this effect was dependent on MHC:TCR interaction, we cocultured CD4 + T cells with TMD8 cells transfected with two siIRF4 or siCtr in presence of blocking antibodies for MHC class I, MHC class II or both. Blocking one or the other MHC class I or II partially rescued the IRF4 effect on T reg and T FH cells respectively as well as reduced anti-lymphoma cytotoxicity. Further experiments blocking IRF4-dependent IL6 and IL10 identified these two cytokines as responsible for the additional effects on T cells. To confirm this as an antigen-dependent immune response we coculture sorted GC B cells from NP-OVA-immunized Cγ 1cre; irf4 fl/fl, Cγ 1cre; irf4 fl/+ and WT mice with CSFE-labeled OVA-specific OT-II CD4 + T cells at a ratio of 2:1 in an organoid system for 5 days. We observed induced antigen specific proliferation of CD4 + T cells in the presence of Cγ 1cre; irf4 fl/fl and Cγ 1cre; irf4 fl/+ GC B cellscompared to WT,as indicated by decrease in CSFE expression . We confirmed an increase of T FH cells and a decrease of T reg cells, supporting the conclusion that IRF4 directly controls the antigen-dependent immune response . In summary, IRF4 expression in GC and lymphoma B cells regulates immune signaling with PU.1 and E2A, thereby modulating antigen-dependent immune responses with potential implications for lymphomagenesis. Future studies investigating targeting of IRF4 in lymphoma are warranted.

Objective: This research delineates the temporal alterations of cell-free circulating tumor DNA (ctDNA) in the plasma and cerebrospinal fluid (CSF) of patients harboring primary and secondary central nervous system (CNS) lymphoma. The aim is to establish that CSF serves as a superior reservoir of ctDNA compared to plasma, thereby aiding in diagnostic processes and monitoring therapeutic efficacy. Concurrently, this study underscores the enhanced response and survival rates in patients administered Bruton's tyrosine kinase (BTK) inhibitors over an extended period exceeding three months. Methods: Data including cranial-enhanced magnetic resonance imaging (MRI) scans, leukocyte and protein concentrations in the CSF, exfoliated CSF cytology, flow cytometry, interleukin-10 to interleukin-6 ratios, ctDNA in plasma and CSF, were gathered from patients presenting primary and secondary CNS lymphoma（7 PCNSL,9 SCNSL）. The temporal ctDNA modifications in plasma and CSF during the patient's disease progression-initial onset, remission, and relapse-were scrutinized. Additionally, the implications of BTK inhibitor usage on patient response rates and survival rates were assessed. Results: Within the cohort of primary CNS lymphoma patients, ctDNA was detected in the CSF and plasma of 2 and 5 patients, respectively. Conversely, in patients with secondary CNS lymphoma, CSF and plasma ctDNA were discovered in 8 and 4 cases, respectively. The detection rate was contingent upon the central lesion's magnitude and position and the patient's surgical history related to the brain. Four patients, including two from each group, underwent cranial-enhanced MRI to appraise tumor burden at various disease stages, simultaneously with the monitoring of ctDNA in plasma and CSF. It was revealed that ctDNA levels in the CSF were in synchrony with the tumor burden, whereas plasma ctDNA levels remained relatively static. This indicates that CSF ctDNA may serve as a viable tool to assess therapeutic responses in CNS lymphoma patients. In two patients where surgical biopsy was unfeasible and both CSF exfoliation cytology and flow cytometry returned negative results, CSF ctDNA was positive, thereby underscoring its auxiliary role in diagnosing CNS lymphoma. Furthermore, the study indicated improved response and survival rates in patients subjected to BTK inhibitors for over three months. Conclusion: CSF ctDNA demonstrates potential in detecting CNS lymphoma and in the dynamic monitoring of therapeutic responses. Compared to cranial-enhanced MRI, exfoliated CSF cytology, and flow cytometry, CSF ctDNA testing offers superior sensitivity. For patients for whom brain surgical biopsy is not an option, CSF ctDNA testing emerges as a valuable supplementary diagnostic tool. CSF ctDNA allows for dynamic monitoring of therapeutic responses in CNS lymphoma patients, enabling the detection of residual lesions, and thereby guiding the formulation and duration of the treatment regimen. Patients administered oral BTK inhibitors for more than three months demonstrate higher treatment response and survival rates.

The single-cell platform mass cytometry (CyTOF) enables detection of significant changes in one million of renal cells, especially in subsets of immune cells in ischemic human kidneys distal to renal artery stenosis (RAS). We found that pro-repair cell types such as scattered tubular-progenitor-like cells, aged neutrophils, and regulatory B cells show a compensatory increase in RAS. Immune cell phenotype changes may reflect ongoing inflammation and impaired immune defense capability in the kidneys.

Abstract T follicular helper (TFH) cells are a subset of CD4+ T cells that reside within the follicles of secondary lymphoid organs and play an essential role in the germinal center (GC) reaction. To examine the role TFH cells in follicular lymphoma (FL), we first characterized TFH cells using single cell analysis with the high-throughput technologies of mass cytometry (CyTOF) and the Indexing of Transcriptomes and Epitopes by Sequencing (CITE-seq). Clustering analysis revealed that TFH cells form subsets with unique phenotypes, and the expression of many common markers including CXCR5, ICOS, IL-21, TOX2, CD57, GZMK, XCL1, XCL2, CXCL13 and CD40L varies on TFH subsets. We found that TFH cells are phenotypically more heterogenous in FL than benign tonsil tissues by both proteomic (n=82) and transcriptomic (n=8) profiling. Each FL patient showed different dominant TFH subsets, in contrast to tonsil donors which were highly homogenous. Some TFH subsets were only present in FL and absent in tonsil tissues, suggesting the existence of FL-specific TFH cells. Cytokines (IL-2, IL12, IL-17 and IL-21) and activated B cells upregulated TFH markers including PD-1, CXCR5 and IL-21 to induce the development of TFH subsets such as IL-21+ or IL-21+IFN-g+ TFH cells, which was inhibited by rapamycin, an immunosuppressor. We observed that the TFH cell frequency significantly correlated with the number of lymphoma B cells. Futhermore, while TFH cells as a whole showed no correlation with prognosis, TFH subsets such as CD57+ TFH cells were significantly associated with an inferior patient outcome. Taken together, our results indicate that TFH cells are highly heterogenous with different tumor-specific TFH subsets present in FL and associated with patient outcomes in FL.

Abstract The tumor microenvironment (TME) plays a crucial role in mediating the tumor immune response, thereby affecting patient outcomes in follicular lymphoma (FL). Using CyTOF, we analyzed a cohort of 82 FL patients with biopsy specimens collected before the first treatment regimen was administered. Our results showed that there was a difference in the T-cell phenotype in patients with various TMEs. Patients with TMEs rich in monocyte/macrophages/NK cells had more CD8+ cells and tended to have fewer TN, but more TEMRA and TEXH cells. Patients enriched with T cells exhibited a T-cell phenotype skewed toward to TCM and patients with B cell dominance tended to have significant higher number of TFH cells. T-cell phenotypes with more terminally differentiated, senescent and exhausted cells were highly represented in patients with less T cells when compared to patients with more T cells. We observed that CD57+ TFH cells was significantly more abundant in patients who had disease progression than patients who had complete response to therapy, suggesting a role for CD57+ TFH cells in promoting malignant cell growth in FL. Consistent with this finding, we observed that increased numbers of CD57+ TFH cells correlated with an inferior survival in FL. Using CITE-seq technology, we found that CD57+ TFH cells exhibited a substantially different transcriptome from CD57− TFH cells. Genes that were differentially upregulated in CD57+ TFH cells when compared to CD57− TFH cells included genes involved in cell survival, compromised inflammatory response, and metabolic activation. Taken together, our results indicate different tumor microenvironments among patient groups that is associated with variable T-cell phenotypes.