Dong, Haidong

T cell exhaustion is often viewed as a terminal state marked by irreversible loss of function during chronic infection and cancer. However, the rapid recovery of antitumor activity in clinical responders after PD-1 blockade suggests that some exhausted T cells retain hidden functional potential. To explain this paradox, we present a theoretical framework that separates latent effector capacity from active effector function. Latent effector capacity represents a history-dependent state reflecting preserved epigenetic accessibility and regulatory readiness, distinct from ongoing gene expression. In this model, PD-1 signaling acts as a reversible mask that suppresses effector function without eliminating the underlying effector capacity. The framework explains preserved chromatin accessibility, rapid reinvigoration after checkpoint blockade, and variable therapeutic responses. By incorporating nonlinear maintenance and erosion of latent effector capacity, the model predicts a bistable system with a point of no return beyond which exhaustion becomes irreversible. These principles may guide development of mechanistically informed biomarkers and artificial intelligence models predicting response to PD-1 therapy.

Abstract Effector gene expression in CD8+ T cells is tightly controlled at the epigenetic and transcriptional levels; however, crucial regulatory signals remain incompletely understood. In this study, we identify dual specificity phosphatase 2 (DUSP2) as a key epigenetic modulator of effector CD8+ T-cell state, by elucidating inhibition of BAF-mediated chromatin opening by DUSP2. Induced downregulation of DUSP2 in human CD8+ T cells enhances cytotoxicity and effector phenotypes by increasing chromatin accessibility through upregulation of BAF subunits, such as SMARCA4, and permitting binding of ETS1 to BAF subunit gene loci and those essential for effector programs. These findings uncover a previously unrecognized role of DUSP2 in the management of the chromatin landscape in effector CD8+ T cells and the epigenetic fine-tuning of their functionality, highlighting DUSP2 as a promising therapeutic target for T cell–based immunotherapy.

Abstract Understanding how immune checkpoint inhibitors (ICIs) reshape human T-cell responses requires models that move beyond static transcriptomic snapshots and discrete cell-state classifications. Here, we present an integrative AI-enabled virtual cell framework that represents human CD8⁺ T-cell responses as dynamic and computable systems during ICI therapy. By integrating single-cell RNA sequencing with paired T-cell receptor sequencing within the C2S-scale foundation model, we construct a virtual representation of individual T cells, in which each cell is encoded by a unique functional identity that captures its transcriptional, signaling, and clonal characteristics. Using this framework, we identify a previously unrecognized dynamic latent effector state of CD8⁺ T cells characterized by intermediate expression of effector genes, distinct signaling activity, and ongoing clonal expansion. Across independent patient cohorts, the virtual cell model consistently indicates that ICI therapy mainly acts by unmasking pre-existing effector potential rather than inducing de novo effector differentiation. Notably, this latent effector population remains transcriptionally restrained despite active signaling and clonal expansion, revealing a hidden reservoir of antitumor immune capacity. More broadly, our study demonstrates how AI-enabled virtual cell modeling can reconstruct latent cellular states and their dynamic transitions from multidimensional single-cell data. By incorporating functional identity into virtual cell model, this framework uncovers biologically meaningful yet non-obvious T-cell effector program during cancer immunotherapy and provides a generalizable approach for studying immune dynamics in human disease.

Cytotoxic T cells produce the cytokine interferon-γ (IFN-γ). We sought to determine whether IFN-γ directly kills target cells and found that T cells used IFN-γ to kill murine melanoma cells lacking the kinases TBK1 and IKKε. In the absence of both kinases, IFN-γ induced the production of TNF receptor 1 (TNFR1) and the sensor Z-DNA binding protein 1 (ZBP1) in the tumor cells to stimulate receptor-interacting protein kinase 1 (RIPK1)–dependent apoptosis in a cell-autonomous manner. IFN-γ also enhanced the activation of nuclear factor κB (NF-κB) signaling in a TNFR1-dependent manner in cells deficient in both TBK1 and IKKε. Because IFN-γ–induced apoptosis occurred in a transcription-dependent manner with slow kinetics, STAT1 and NF-κB cooperated to activate the expression of inflammatory genes in the dying cells. Thus, IFN-γ–induced cell death was accompanied by an inflammatory signature in the absence of TBK1 and IKKε. TBK1 and IKKε not only mediated the induction of type I IFN but also inhibited RIPK1-dependent death and NF-κB–dependent inflammation. Therefore, we propose that these kinases may have gained these intertwined functions so that cells infected by pathogens that produce antagonists of TBK1 and IKKε are eliminated by IFN-γ–secreting T cells in an inflammatory manner to compensate for the inhibition of type I IFN production.

Our research uncovers a role for ATR in responding to ECM stiffness and promoting epithelial-to-mesenchymal transition (EMT) and metastasis. ATR, when deubiquitinated and upregulated by USP21 under enhanced ECM stiffness conditions, phosphorylates the nuclear protein SUN2, which promotes β-catenin nuclear translocation and EMT. ATM-mediated EMT promotes polymorphonuclear myeloid-derived suppressor cell recruitment and inhibits CD103 + dendritic cells, fostering an immunosuppressive tumor milieu. ATR inhibition disrupts this malignant cascade by promoting mesenchymal-to-epithelial transition to enhance antitumor immunity and mitigate metastases. Consistently, circulating HLA-DR + dendritic cells were also enhanced following treatment with the ATR inhibitor berzosertib in patients with therapeutically resistant early-stage breast cancer. Our data suggest that ATR-targeted therapy may be optimized by considering both DNA damage–dependent and EMT-inducing effects of ATR ATR promotes metastasis and immunosuppression through EMT process

Abstract Macrophages in the tumor microenvironment are known to upregulate PD-L1 expression, thereby suppressing T cells through PD-1 ligation. However, the manner in which PD-L1 expression intrinsically impacts macrophages and their immunomodulatory phenotype is less clear. Clarifying this knowledge gap would yield insight into the mechanisms of immunosuppression within the tumor microenvironment. To characterize the macrophage intrinsic role of PD-L1, we used complementary genetic and pharmacological approaches by analyzing primary murine bone marrow-derived macrophages (BMDMs) with complete genetic PD-L1 deletion and wildtype BMDMs treated with anti-PD-L1 blocking antibodies. Macrophages were evaluated across naïve, pro-inflammatory (M1), and tumor conditioned (TCM) polarization states in vitro. Unlike prior reports, neither genetic deletion nor antibody blockade dramatically altered the expression of macrophage polarization markers or in vitro phagocytic capacity. Both conditions consistently reduced surface levels of the M1-associated costimulatory molecule CD80, prompting further analysis of T cell interacting and antigen presenting proteins, in which we revealed disparate effects of genetic deletion and antibody blockade on the surface levels of MHCI, MHCII, PD-1, and PD-L2. These findings suggest that PD-L1 deletion and antibody-mediated blockade contribute to macrophage immune regulatory profiles in distinct manners. This difference supports a model in which PD-L1 functions in macrophages beyond its canonical role as a ligand for PD-1, influencing antigen presentation and checkpoint molecule levels and playing a broader role in immune regulation in the tumor microenvironment.

ABSTRACT A central paradox in cancer immunotherapy is that patients harboring apparently exhausted or dysfunctional CD8 + T cells can exhibit rapid and durable responses to immune checkpoint inhibition (ICI). Although such rebound responses do not occur in all ICI-treated patients, resolving this paradox is essential for identifying and therapeutically maximizing reversible T cell exhaustion before clinical benefit is achieved. Here, we identify malic enzyme 1 (ME1) as a key molecular determinant of a latent, epigenetically poised effector CD8 + T cell state that is both necessary and sufficient for responsiveness to ICI. Genetic loss of ME1 abolishes therapeutic efficacy despite intact checkpoint blockade, whereas enforced ME1 expression enables robust antitumor responses even in otherwise ICI-resistant tumor models. Chromatin accessibility profiling, together with bulk and single-cell transcriptomic analyses, demonstrates that ME1 preserves effector readiness by maintaining latent effector capacity as a measurable biological state. These findings also experimentally ground a mechanistically interpretable dynamical model of T cell exhaustion, in which latent effector capacity, E(t), evolves according to a first-order differential equation shaped by activation history and decay. By defining latent effector capacity as a quantifiable, history-dependent state variable, our findings further enable the rational design of AI-driven models to predict patient-specific responsiveness and optimize therapeutic strategies in T cell–based immunotherapy. Together, our results redefine exhaustion as a state of masked potential governed by history-dependent latent effector capacity and provide a unified framework explaining rapid functional rebound, therapeutic heterogeneity, and the boundary between reversible and irreversible T cell exhaustion.

Abstract T cell exhaustion is commonly viewed as a terminal differentiation state marked by irreversible loss of effector function during chronic infection and cancer. However, the rapid restoration of cytotoxic activity following PD-1 checkpoint blockade challenges this view, revealing a central paradox: T cells that appear functionally inert can regain effector function on timescales incompatible with de novo differentiation or extensive epigenetic reprogramming. To resolve this contradiction, we present a mathematical framework that explicitly decouples latent effector capacity from active effector output. We define latent effector capacity as a slow, history-dependent state variable representing preserved epigenetic accessibility and regulatory readiness at effector loci, distinct from instantaneous transcriptional activity. Within this framework, PD-1 signaling functions as a reversible, graded masking mechanism that suppresses effector realization without erasing latent capacity, thereby explaining the coexistence of preserved chromatin accessibility, rapid functional rebound, and heterogeneous responses to checkpoint blockade. Incorporating nonlinear self-maintenance of epigenetic programs together with checkpoint-dependent erosion of latent capacity reveals a bistable regime and a history-dependent point of no return, beyond which exhaustion becomes irreversible. Critically, the model demonstrates that PD-1 checkpoint blockade unmasks pre-existing effector potential but cannot recreate lost capacity, because therapeutic reversibility is governed by the prior dynamical stability of a latent epigenetic state rather than by instantaneous transcriptional output. This framework establishes a mathematical foundation for mechanistically predictive AI in PD-1 blockade therapy by identifying latent, history-dependent variables that can be inferred from epigenetic and transcriptional data to predict therapeutic responsiveness and irreversibility.

The underlying immunopathogenesis of inflammatory arthritis (IA) immune-related adverse event (irAE) remains obscure. Unlike rheumatoid arthritis (RA), where autoantibodies and B cell dysfunction are central features, the contribution of humoral immunity to IA-irAE is unclear. Here, we performed immunophenotyping of peripheral blood from patients with IA-irAE and compared them with patients with seronegative RA, immune checkpoint inhibition–treated patients without irAE, and healthy controls. IA-irAE was marked with increased cytotoxic gene expression and metabolic activation in T cells and reduced CXCR3 and CCR6 expression in CD4 + T cells. Contrary to seronegative RA, patients with IA-irAE displayed no substantial elevation in autoantibody levels or atypical CD11c + CD21 − B cells. IA-irAE was further characterized by elevated levels of interleukin-6 (IL-6), IL-12, and type I interferon, which correlated with the T cell activation phenotypes. Together, our findings define IA-irAE as a disease with certain immunological features distinctive from RA, representing a potentially T cell–driven, autoantibody-independent autoimmunity. These results offer insights into immune tolerance breakdown and therapeutic targeting in irAEs.

Background Immune checkpoint therapies targeting programmed cell death protein 1 (PD-1) and programmed death ligand 1 (PD-L1) have been met with limited clinical responses in the treatment of patients of varying cancer types. Current US Food and Drug Administration (FDA)-approved therapies function to prevent PD-1 binding to PD-L1, thereby preventing T-cell suppression. However, these therapies fail to prevent PD-L1 intrinsic signaling and recycling, which can render these treatments ineffective. Moreover, most research primarily focuses on how these antibodies affect tumor cells and their interactions with T cells. Myeloid cells, which also express PD-L1 and influence T-cell responses, are integral to the efficacy of these therapies. However, strategies that can target these cells to boost antitumor responses have not been effective. Methods Our group has developed a new anti-PD-L1 antibody (H1A), which disrupts PD-L1 recycling and redirects it towards degradation. Due to the loss of available PD-L1 protein, H1A effectively prevents PD-L1 intrinsic signaling. Using in vivo humanized PD-1/PD-L1 mouse tumor models, we evaluated the therapeutic efficacy of H1A against current FDA-approved PD-L1 blocking antibodies. Using human-derived immune cell in vitro assays, we pinpoint the biological consequences of H1A-induced PD-L1 degradation in myeloid cells. Results H1A demonstrated improved tumor control and established immunological memory responses in vivo in the treatment of tumors with moderate immunogenicity that are less responsive to current immunotherapies. This was supported by enhanced activation of myeloid cells (major histocompatibility complex (MHC)-II and CD80) and frequencies of effector T-cell populations found intratumorally. Human myeloid cells treated with H1A were also observed to have increased activation (MHC-II and CD80) and cytokine secretion. Similar to our in vivo findings, human peripheral blood lymphocyte cultures demonstrated increased frequencies of effector T-cell populations and greater tumor cell killing. Conclusions This work highlights the importance of targeting PD-L1 recycling and intrinsic signaling, specifically in myeloid cells, to boost antitumor responses. Our studies suggest that H1A can provide a solution to the lack of responses seen with current therapeutics, while also revealing previously unknown intrinsic functions of PD-L1 in myeloid cells.