Figure S1. Cellular composition of the post-treatment tumor biopsy of patient UPN-03. Figure S2. Gating strategy for multiparameter flow cytometry for UPN-03. Figure S3. Gating strategy for multiparameter flow cytometry for UPN-04.

ABSTRACT:Autologous stem cell transplantation (ASCT) with maintenance lenalidomide remains the mainstay of consolidation therapy for eligible patients with multiple myeloma, but preventing disease relapse remains a critical unmet need. Here, we investigated whether immunosuppressive myeloid populations in bone marrow correlated with ASCT outcomes. We identified a subset of CD64+CD169+CD163+ macrophages that expressed CSF-1R, PD-L1, and CD155 and were expanded in patients who relapsed after ASCT. Using a preclinical ASCT model with suboptimal endogenous antimyeloma activity, we demonstrated that although neither CSF-1R inhibition nor lenalidomide monotherapy significantly improved outcomes, their combination synergistically attenuated disease progression and prolonged survival. Single-cell RNA sequencing revealed that lenalidomide expanded natural killer (NK)-like CD8+ T cells but paradoxically also increased the frequency of Csf1r+ macrophages. Cell-cell communication analyses identified Csf1r+ macrophages as suppressors of these NK- and effector-like phenotypically exhausted (Tphex) CD8 T-cell populations through CD94/NKG2A and PD-L1/PD-1, respectively. CSF-1R blockade depleted these immunosuppressive macrophages, which correlated with decreased expression of inhibitory receptors and enhanced expression of activation markers in Tphex. Given the US Food and Drug Administration approval of axatilimab for chronic graft-versus-host disease, combining CSF-1R blockade with lenalidomide maintenance represents a readily testable strategy to improve progression-free survival after ASCT.

Antibody-secreting cells (ASCs) provide humoral immunity that can mediate lifelong protection against pathogens. Current classifications cannot delineate the heterogenous functionalities, tissue residencies, and lifespans of human ASC subsets, impeding clinical translation. We applied multi-omic sequencing, spatial proteomics, and functional assays to discover and characterize human bone marrow (BM) ASC subsets. We identified two peripheral subsets (ASCp) also present in blood and three BM-resident subsets (ASCr), comprising a maturation continuum associated with increased mitochondrial networking, diminished antibody secretion, differential transcription factor motif accessibility, and preferential co-localization in homotypic niches. CD19+9+ASCr and CD19-ASCr exhibited poor recovery years after BM transplantation, indicating a strong dependence on supportive niches. Childhood vaccine antigens were recognized by long-lived ASCr subsets in adults and by immature HLA-DR+ASCp, implying ASCs can differentiate without recent antigen exposure. Our results provide new insights into ASC identity, maturation, and longevity and a generalizable framework for study and manipulation of human ASCs.

Background:Chimeric Antigen Receptor (CAR) T cell therapy targeting B-cell maturation antigen (BCMA) has demonstrated impressive clinical efficacy in relapsed/refractory multiple myeloma (MM). Nonetheless, disease relapse limits durable response for most patients. Trogocytosis of target antigen by effector cells has emerged as a potential contributor to reduced surface antigen density, CAR T cell dysfunction, and fratricide. Although γ-secretase inhibitors (GSI) significantly increase cell surface BCMA density and decrease soluble BCMA (sBCMA), their effects on BCMA trogocytosis and the resulting impact on CAR T-cell function remain incompletely understood. Methods:We investigated the effects of GSI on BCMA-directed CAR T cell function and trogocytosis using in vitro co-culture systems with MM cell lines across a spectrum of BCMA expression. We validated findings using confocal microscopy and cytotoxicity assays. Trogocytosis and fratricide were assessed in time-resolved functional studies. Phenotypic and functional differences between trogocytosis-positive (CAR T Trogo+) and trogocytosis-negative (CAR T) cells were evaluated using multiparametric flow cytometry, proteomic profiling, single-cell RNA sequencing (scRNA-seq), T-cell receptor (TCR) sequencing, and in vitro rechallenge assays. We also interrogated clinical samples from two Phase I trials (NCT03338972 and NCT03502577) which employed the identical CAR T cell construct with or without GSI respectively, to evaluate the relationship between trogocytosis, CAR T cell persistence, and treatment outcome. Results:GSI driven increases in BCMA density on MM cell lines enhanced CAR T cell cytotoxicity but concomitantly increased trogocytosis, particularly in high-antigen-density cell lines-(H929+GSI vs H929; 30 min (P<0.0001), 1 h (P<0.0001), 2 h (P<0.0001), 6 h (P<0.0001), and 24 h (P<0.0001) and in CD4+ CAR T cells (K562mCherry+GSI, CD4 vs CD8 CAR T cells;10 min (P=0.01), 2 h (P=0.01), and 6 h (P=0.004). Following BCMA acquisition, CAR T cells (CAR T Trogo+) exhibited reduced proliferative capacity, diminished cytotoxic function (CAR T Trogo+ vs CAR T; (P=0.01), and an increase in markers of exhaustion/activation (CD4+ CAR T Trogo+ vs CD4 CAR T and CD8+ CAR T Trogo+ vs CAR T; PD-1+LAG-3+, PD-1+TIM-3+, and TOX+TIM-3 co-expression, (P=0.007, P<0.0001, P=0.006 and P=0.004, P=0.03, P=0.006). In fratricide assays, CAR T Trogo+ cells were susceptible to killing by naïve CAR T cells. Single cell RNA-seq supports the phenotypic findings revealing transcriptional features of heightened activation and accelerated exhaustion in CAR T Trogo+ cells. Clinical phase I trial data confirm BCMA trogocytosis in patient samples. Conclusions:Our findings highlight the paradoxical effects of increased BCMA density on BCMA CAR T cell therapy: enhancement of initial tumor targeting and promotion of trogocytosis-associated dysfunction. Trogocytosis may contribute to antigen modulation, CAR T cell exhaustion, and fratricide, potentially muting the therapeutic benefits of enhanced antigen density. To optimize GSI and mitigate trogocytosis-associated resistance mechanism, future clinical trial designs should incorporate early time-point sampling, a sample size providing sufficient statistical power to determine an impact on CAR T cell persistence and treatment response, and mechanistic assessments.

Humoral immune-related adverse events, including hypogammaglobulinemia and B cell depletion, pose long-term infection risks after chimeric antigen receptor T cell therapy (CARTx) for hematologic malignancies. This prospective study evaluates the kinetics of pathogen-specific humoral immunity prior to and up to a year after CARTx targeting CD19 and CD20 (B cells) or BCMA (plasma cells) in 100 and 28 individuals, respectively. Antibodies are tested for 12 vaccine-preventable pathogens and using comprehensive high-throughput antibody profiling. A subset of 72 participants are evaluated for post-CARTx vaccine responses. Here, we show pathogen-specific humoral immunity does not significantly change after CD19-, CD20-, or BCMA-targeted CAR-T cell therapy (CARTx). However, seroprotective antibodies are absent for up to one-third of routine vaccine-preventable pathogens in CD19- and CD20-CARTx recipients and for nearly half of vaccine-preventable pathogens in BCMA-CARTx recipients by one-year post-CARTx. Pre-vaccination B cell count is the main predictor of vaccine response.

Treatment-refractory and relapsed disease remain leading causes of death for patients with lymphoma. Virtually all lymphomas are exquisitely sensitive to radiation, and α-particle radiation therapies are notably suited to targeting microcluster disease common in the setting of early relapse. Refractory or relapsed lymphoma may also involve the loss of therapeutic targets, but radiation may stimulate antitumor immune effects against disease with incomplete target expression. Such effects make immune checkpoint inhibition a compelling candidate for combination treatment. Methods: We evaluated the therapeutic efficacy of 211At-labeled antihuman CD20 monoclonal antibodies combined with immune checkpoint inhibition in human CD20 transgenic mice bearing murine lymphomas on opposing flanks that were either positive or negative for human CD20 expression (hCD20(+) and hCD20(-), respectively). Results: In the absence of 211At-hCD20, the antimurine checkpoint inhibitors PD1, CTLA4, CD47, and TIM3 had no efficacy given alone or in doublets. 211At-hCD20 given alone suppressed growth of both hCD20(+) and hCD20(-) tumors in a dose-dependent fashion, with predictably stronger suppression of hCD20(+) tumors. Strikingly, the addition of PD1 alone or the PD1 plus CTLA4 doublet to low-dose 211At-hCD20 significantly strengthened suppression of both tumors and increased mouse survival. Conclusion: Future translation of this synergistic combination of α-radiotherapy and immune checkpoint inhibition holds promise for the treatment of high-risk aggressive lymphomas, including cases with postinduction minimal residual disease or antigen loss after targeted therapies.

Background/Objectives: Autologous stem cell transplantation (ASCT) remains a standard frontline treatment for eligible patients with multiple myeloma and is associated with im-proved progression-free survival and sustained treatment responses. In addition to medi-cal eligibility factors, pre-treatment psychosocial variables such as depression and anxiety as well as cognitive functioning may influence short and long-term transplant outcomes. This study examined the relationship between pre-transplant psychosocial and cognitive functioning and clinical outcomes in patients with multiple myeloma undergoing ASCT. Methods: Participants (N = 245) undergoing ASCT in Miami, Florida completed routine pre-transplant psychosocial and neuropsychological assessments. Patient-reported clini-cal variables and cognitive performance were examined in relation to transplant success and transplant recovery outcomes, including time to engraftment, length of hospital stay post-engraftment, and survival across four timepoints. Results: While most patients achieved engraftment within a similar timeframe, the duration of hospitalization follow-ing engraftment varied considerably. Higher baseline trait anxiety was associated with longer post-engraftment hospital stays. In multivariable Cox regression analyses, post-engraftment length of stay was independently associated with long-term mortality when modeled as a continuous variable. In a secondary analysis using an empirically de-rived cutoff corresponding to stays longer than the 80th percentile (≥7 days post-engraftment), both prolonged post-engraftment hospitalization and poorer baseline delayed memory were independently associated with mortality. Descriptive survival analyses further showed that patients with prolonged post-engraftment hospitalizations had higher overall mortality and shorter survival than those with shorter post-engraftment hospital stays. Conclusions: Patients who demonstrated delayed memory performance pre-transplant and prolonged hospital admission post-transplant may be most vulnerable to dying when examined at long-term follow-up. Pre-transplant psychiatric assessments provide an opportunity to identify potentially implement supportive care interventions for patients at increased vulnerability for cognitive impairment.

Table S3. Cellular composition of post-infusion timepoints.

Introduction Prolonged thrombocytopenia (high-grade thrombocytopenia lasting more than 28 to 30 days after CAR T-cell therapy) affects approximately 22% to 54% of patients (Brudno JN, et al JAMA. 2024;332(22):1924–1935). This study analyzes predictors of prolonged thrombocytopenia following CAR-T therapy. Objectives 1. To determine predictive factors related to post CAR-T thrombocytopenia2. To identify malignant cellularity within the bone marrow pre CAR-T therapy as a predictor of post CAR-T thrombocytopenia3. To acknowledge thrombocytopenia post CAR-T is secondary to a hypoproductive process as demonstrated by low immature platelet fraction Methods 284 patients underwent CAR-T therapy at our institution from 2016 to 2025 (lymphoma or lymphocytic leukemia (N=257), myeloma (n=27)). We analyzed the immature platelet fraction (IPF) in 226 patients with lymphoma or lymphocytic leukemia. We also assessed the percentage of total and malignant cellularity within the latest pre-treatment bone marrow biopsy, available in 134 patients with lymphoma, lymphocytic leukemia, or myeloma. Results Prolonged thrombocytopenia (<25 × 10^9/L) was identified in 32% patients. Baseline thrombocytopenia and marrow infiltration with malignant cells were the two parameters that predicted prolonged thrombocytopenia at 30 days. At 30 days: patients with severe thrombocytopenia had mean baseline platelets of 85 × 10^9/L (95% CI 79-90) versus 166 × 10^9/L (95% CI 169-172) without severe thrombocytopenia. Overall bone marrow cellularity (pre-CAR-T) did not correlate with prolonged thrombocytopenia at 30 days. As cellularity reflects a mixture of hematopoietic and malignant cells, we assessed the predictive value of percent malignant cells within the bone marrow.Malignant cell infiltration in baseline bone marrow biopsy predicted prolonged thrombocytopenia (malignant involvement: <=5% malignant involvement, mean platelets 62.5 × 10^9/L (95% CI 55.3-69.6), 6-49% malignant involvement, mean platelets 71.2 × 10^9/L (95% CI 57.1-85.3), >=50% malignant involvement, mean platelets 41.2 × 10^9/L (95% CI 31.6-50.7).At day 30, the mean IPF was not significantly increased in the patients with severe thrombocytopenia (5.8% (CI 4.1-7.6)) compared to those without severe thrombocytopenia (4.6% (CI 3.4-5.8)). Conclusion We confirmed the high incidence of prolonged thrombocytopenia after CAR-T therapy. Malignant involvement of the bone marrow (>=50%) at baseline, and baseline thrombocytopenia, both predicted prolonged thrombocytopenia. The non-elevated IPF further supports the hypoproductive mechanism, as IPF is elevated in patients with increased peripheral destruction of platelets. (Goel G, et al, J Lab Physicians, 2021).As the two predictors of prolonged thrombocytopenia are available pre-treatment, these findings may identify a subgroup for whom thrombopoietic support may be studied.

Table S2. Differentially expressed TCR-β CDR3 sequences increased at 3 and/or 35 month timepoints in patient UPN-04

Advances in immunotherapy for acute myeloid leukemia (AML) have revealed critical gaps in selection of appropriate preclinical models. Conventional cytotoxic and targeted therapies could be tested in relatively straightforward systems. Immunotherapies are inherently distinct from conventional therapies, as their activity is shaped by dynamic, context-dependent interactions between leukemic cells, immune effectors, and the bone marrow microenvironment that are difficult to replicate outside the patient. Failure to recapitulate these interactions may limit the clinical translatability of promising results. Recognizing these shortcomings, the field has moved through several generations of modeling platforms. Existing platforms capture distinct but incomplete aspects of AML biology: in vitro systems offer control but lack immune context, syngeneic models provide intact immunity but limited human relevance, patient-derived xenograft (PDX) models preserve patient biology but lack immunity, and humanized models partially integrate both but remain constrained. In this review, we trace the development of these systems and use that trajectory to build a practical framework for model selection in AML immunotherapy research. Translational relevance depends on selecting preclinical models that align with the specific therapeutic question, rather than relying on availability or perceived complexity. To operationalize this, we propose a mechanism-driven framework that maps therapeutic mechanisms to the most appropriate biological contexts across preclinical platforms. This framework progresses sequentially from controlled mechanistic studies to in vivo validation, guiding model selection at each stage. Instead of prioritizing any single system, it emphasizes a complementary, question-driven approach that leverages the distinct strengths of each model while accounting for their limitations.

Supp Figure SF1 shows Treatment Schema

PK parameter estimates for the nirogacestat dose-effect model

Nirogacestat treatment decreases sBCMA concentration

Participant demographics and baseline characteristics

BACKGROUND:BCMA-directed chimeric antigen receptor (CAR)-T cell therapy represents a major therapeutic breakthrough for relapsed/refractory multiple myeloma (RRMM), offering deep and durable responses in heavily pretreated patients. However, a subset of patients experiences early relapse or fail to respond, highlighting the need for strategies to enhance efficacy. Gamma-secretase inhibitors (GSIs) have been shown to increase surface BCMA expression on malignant plasma cells and may potentiate the activity of BCMA CAR-T cells, particularly in patients with low baseline BCMA antigen density. Two contemporaneous phase 1 trials (FH9952 and FH9762) evaluated the fully human BCMA-targeted CAR-T product FCARH143, with FH9952 incorporating GSI co-administration. OBJECTIVE:To determine whether GSI use improves clinical outcomes following FCARH143 CAR-T therapy in RRMM, with particular focus on overall survival (OS), progression-free survival (PFS), and subgroup differences based on prior BCMA-targeted therapy and baseline tumor BCMA levels. STUDY DESIGN:This retrospective analysis compared outcomes from two single-arm, single-center phase 1 trials of the fully human BCMA-targeted CAR-T cell product FCARH143 in RRMM: FH9952 (n = 18), which incorporated the GSI crenigacestat, and FH9762 (n = 25), which did not. Both studies had extended follow-up beyond previously published reports. Eligible patients had measurable RRMM with ≥10% bone marrow plasma cell involvement and confirmed BCMA expression. BCMA-naïve patients were defined as those without prior exposure to BCMA-targeted therapies (e.g., BCMA CAR-T, bispecific antibodies, or antibody-drug conjugates); BCMA-exposed patients had received at least one such therapy. All participants received lymphodepletion with fludarabine and cyclophosphamide followed by infusion of FCARH143. In FH9952, crenigacestat (25 mg orally, three times weekly) was administered from day 0 through day +18. Outcomes included response rates, adverse events, OS, and PFS. Cox proportional hazards models were used for survival analysis. RESULTS:With extended follow-up (median 5.8 years), the overall cohort (n = 43) had a median OS of 2.7 years (95% CI 1.9-4.8) and median PFS of 1.1 years (95% CI 0.72-2.2). Baseline characteristics were generally similar, though BCMA-exposed status was more frequent in FH9952 (39% versus 8%). Rates of cytokine release syndrome, immune effector cell-associated neurotoxicity, and early immune effector cell-associated hematotoxicity were comparable between trials. Among BCMA-naïve patients (n = 34), GSI use was associated with improved OS (not reached versus 2.3 years; adjusted HR 0.30, 95% CI 0.10-0.88, P = .028) and a trend toward improved PFS (2.6 versus 1.3 years; adjusted HR 0.47, 95% CI 0.22-1.04, P = .062). No benefit was observed among BCMA-exposed patients. Exploratory spline modeling suggested GSI mitigated the adverse impact of low tumor BCMA levels, with inferior outcomes at low BCMA expression observed only in the non-GSI cohort. CONCLUSIONS:Co-administration of a GSI with BCMA CAR-T therapy was associated with improved survival in BCMA-naïve RRMM patients, particularly those with low baseline tumor BCMA levels. These findings suggest GSI modulation of BCMA surface expression may enhance CAR-T efficacy in select patients and support further prospective investigation.

Interferons (IFNs) orchestrate diverse immune responses, but distinguishing individual IFN contributions in human transcriptomic data is challenging due to overlapping interferon-stimulated gene (ISG) signatures and limited cell-type-specific datasets. To address this, we generated a single-cell transcriptomic atlas of IFN responses by stimulating primary human T, B, NK, and CD14 monocytes with IFN-I, IFN-II, and IFN-III. This revealed core and cell-type-specific ISG programs across 13 subsets, highlighting distinct functions of IFNs. We developed an algorithm to separate IFN-I and IFN-II activity in transcriptomic data. Applied to multiple myeloma samples, it showed elevated IFN-I and IFN-II responses, with induction therapy reducing only IFN-I. Extending to multiple disease datasets provided a cross-disease overview of IFN-I and IFN-II activities and revealed increased IFN-II activities in T cells during lupus flares. This resource and the accompanying analytical framework enable dissection of IFN-driven transcriptional programs in a cell-type specific manner in human disease.