NK cell-based therapies for solid tumors are limited by poor persistence and progressive functional impairment. Here, we investigated how the strength of NKG2D stimulation shapes NK cell fate and therapeutic efficacy. Using B10G5, a MICA/B-targeting partial agonist antibody that attenuates NKG2D stimulation, we show that reduced NKG2D stimulation reshapes the intratumoral NK cell compartment by enriching EOMEShiTCF1hi stem-like NK (stemNK) cell pool, preserving functional diversity, and limiting mitochondrial oxidative stress and exhaustion induced by chronic sMIC/MIC exposure. Mechanistically, we identified the H3K27me3 demethylase KDM6B as a conserved regulator of NK cell stemness whose enzymatic activity is required to maintain the intratumoral stemNK cell pool. Conversely, strong NKG2D stimulation induced mitochondrial ROS accumulation and upregulated the inhibitory receptor LAG3 on NK cells. Attenuated NKG2D stimulation on human NK cells increased KDM6B expression and reduced repressive H3K27me3 mark at the TCF7 locus. Importantly, attenuated NKG2D stimulation during ex vivo expansion enhanced human NK cell in vivo persistence and antitumor efficacy following adoptive transfer. Together, these findings establish NKG2D stimulation strength as a key determinant of NK cell fate and function in solid tumors and identify a translatable concept to improve NK cell-based immunotherapy for solid tumors.

To eliminate the persistent reservoir of cells harboring intact HIV-1 proviruses in people living with HIV-1 (PLWH), cure strategies like the Shock-and-Kill approach rely on effector functions of cytolytic T lymphocytes (CTL). CTL are involved in the initial control of HIV-1 viremia and target productively infected cells throughout the course of infection. However, selective killing of susceptible cells could generate a reservoir dominated by cells with dysregulated cell death pathways or other features conferring resistance to killing. Here, we use CTL-engaging single-chain diabodies to assess the rate of lysis of uninfected and HIV-1-infected primary CD4+ T cells under identical CTL pressure in the settings of both latent and active infection. Our findings indicate that with this mode of CTL triggering, infected and uninfected CD4+ T cells from PLWH on ART are generally lysed at identical rates, and that an apparent survival advantage for actively infected CD4+ T cells primarily reflects the reduced surface antigen availability through previously described Nef-dependent downregulation of MHC class I molecules. No survival advantage is observed when the CTL response is directed through diabodies to the stably expressed non-classical MHC class I molecule HLA-E, indicating equal susceptibility to cell death.

ASXL1 mutations confer a poor prognosis in myeloid neoplasms, including in CALR-mutated myeloproliferative neoplasms (MPN). We generated Calr-Asxl1 co-mutant mice and found more severe thrombocytosis and increased splenomegaly, compared with Calr single-mutant animals. Calr-Asxl1 co-mutant megakaryocytes displayed more pronounced morphological abnormalities and enhanced emperipolesis characteristic of inflammation. Concordant with this, a differentiation trajectory directly to megakaryocytes from hematopoietic stem cells (HSC) was more prominent in Calr-Asxl1 co-mutant mice. In this non-canonical trajectory, where megakaryocyte-biased HSCs directly differentiate into megakaryocyte progenitors, we found upregulation of Mef2c, a key transcriptional regulator of megakaryocyte lineage commitment. Mature Calr-Asxl1 co-mutant megakaryocytes displayed an enhanced proinflammatory transcriptional state, whereas the transcriptional signature of Calr-Asxl1 co-mutant HSCs was indicative of inflammation-resistance. Genotyping of transcriptomes (GoT) confirmed an activated megakaryocytic differentiation program in CALR-ASXL1 co-mutant HSCs in primary human MPN bone marrow. These findings link the enhanced severity of CALR-ASXL1 co-mutant MPN to altered megakaryocyte differentiation and help explain some of the biological basis for the more severe MPN phenotype observed in patients.

The establishment of mixed hematopoietic chimerism is a promising way to induce immune tolerance for islet replacement therapy and to treat the underlying autoimmunity in type 1 diabetes (T1D). Mixed chimerism not only promotes effective thymic negative selection of autoreactive cells but also restores regulatory T cell (Treg) function and peripheral tolerance. In the current study, we determined that a novel class of donor-derived CD8+CD44+CD122+ Tregs (d-CD8+CD122+ Tregs) plays a crucial role in controlling autoimmunity in non-obese diabetic (NOD) mice with induced mixed chimerism. Using adoptive T cell transfer experiments, we showed that d-CD8+CD122+ Tregs abrogate autoimmunity by suppressing the exogenously injected diabetogenic T cells in Recombination-Activating Gene deficient NOD mice. These d-CD8+CD122+ Tregs from NOD chimeras show upregulation of Helios, Programmed cell death protein 1, perforin, granzyme-B, CD39, Folate receptor 4, and downregulation of proinflammatory marker SCART1. Using in vitro assays, we show that d-CD8+CD122+ Tregs respond specifically to a Complementarity-Determining Region-3 peptide sequence derived from T cell receptors of islet antigen-specific autoreactive T cells. Similarly, we found that individuals with T1D have a deficiency in CD8+CD122+ Tregs, suggesting a potential loss of regulatory function accompanies disease onset. Revitalizing CD8+CD122+ Tregs may offer a new therapeutic strategy of restoring immune tolerance in autoimmune diabetes.

Golgin-45 (G45) is a coiled-coil protein that transits reversibly between the Golgi and nucleus. In the Golgi, G45 serves as a tether for effectors of a pro-tumorigenic secretory program, but its nuclear role remains unclear. Here, we report that, upon entering the nucleus, G45 functions as a basic leucine zipper transcription factor. From genomic mapping, RNA-sequencing, and functional validation studies, we identify G45-binding sites and show that nuclear G45 can function as a transcriptional activator or repressor, a versatility enabled by phase separation into nuclear bodies enriched in transcriptional co-regulators. In lung cancer models, Golgi-docking reinforces, whereas nuclear translocation opposes, pro-tumorigenic activity of G45. Nuclear G45 incorporates into dimeric transcription factors that silence pro-tumorigenic secreted cargos. Epigenetic events in lung cancer degrade nuclear G45 and sequester G45 in the Golgi, which unleashes malignant secretion. Thus, inter-organelle transitions govern a G45-dependent secretory axis that drives lung cancer progression.

D1-type receptor (D1R)-mediated dopaminergic signaling within the nucleus accumbens shell (NAcSh) is essential for forming adaptive circuit changes that embed persistent memories associated with drug seeking and craving. While D1R is expressed in both NAcSh neurons and astrocytes, the specific contribution of astrocytic D1R to drug-related circuit plasticity remains poorly understood. Here, we demonstrated in mouse NAcSh slices that D1R agonists increased astrocytic Ca2+ activity, an effect that was attenuated by astrocyte-specific D1R knockdown. Furthermore, selective knockdown of astrocytic D1R in the NAcSh prior to cocaine self-administration inhibited cocaine-induced generation of silent synapses, thereby hampering the associated remodeling of NAcSh circuits. Behaviorally, knockdown of NAcSh astrocytic D1R expedited the extinction of cocaine seeking and reduced cue-induced reinstatement. These results identified astrocytic D1R as a fundamental component of NAcSh dopamine signaling during cocaine experience that remodeled synaptic connections and neural networks underlying drug seeking and relapse.

Engineered T and NK cell therapies have widely been used to treat hematologic malignancies and solid tumors, with promising clinical results. Current chimeric antigen receptor (CAR) T cell therapeutics are, however, associated with treatment-related adverse events such as cytokine release syndrome (CRS) and are prone to immunologic exhaustion. CAR-NK therapeutics, while not associated with CRS, have limited in vivo persistence. We now demonstrate that a population of NK-like TCRαβ+ CD8 T cells (TKLR), identified and expanded ex vivo through its expression of the activating receptor NKG2C, can be transduced to express a second-generation CD19 CAR (1928z). These CAR+ TKLR demonstrates superior tumor clearance and longer persistence compared to 1928z CAR+ conventional CD8 T cells (TConv) and 1928z CAR+ NK cells in mouse xenograft models. Moreover, CAR-modified TKLR had decreased expression of PD-1, TIM-3, and LAG-3 and elicited significantly reduced CRS compared to CAR+ CD8 TConv in mouse xenograft models. Similarly, NKG2C+ TKLR engineered with a TCR targeting the NY-ESO-1 antigen exhibited robust tumor control compared to TCR-engineered CD8 TConv in a mouse xenograft model. These data establish NKG2C+ TKLR as a promising platform for cell engineering, and offer a safer, more durable alternative to conventional CAR-T and CAR-NK therapies.

Chronic myelomonocytic leukemia (CMML) is a hematological neoplasm characterized by expansion of malignant monocytes. However, as a stem cell neoplasm, the disease clone comprises most blood lineages, including granulocytes. To investigate their pathogenic contribution, we performed extensive phenotypic, transcriptomic and functional characterization of CMML granulocytes. These cells exhibited reduced granularity and impaired maturation, the latter of which is associated with poor prognosis. Functionally, CMML granulocytes displayed reduced phagocytic capacity and impaired NETosis, together with aberrant gene expression programs. Notably, RETN, encoding the chemokine resistin, was uniquely upregulated in CMML granulocytes. Accordingly, resistin levels were 10-fold higher in CMML plasma, correlating with inferior survival. Ex vivo exposure to resistin directly blocked monocyte maturation and macrophage differentiation, partly via TLR4 interaction, recapitulating the classical monocytosis repartitioning phenotype that characterizes CMML. Resistin-treated monocytes upregulated gene signatures related to immune suppression and myeloid-derived suppressor cell phenotype. Consistent with these findings, CMML patients displayed higher Treg proportions and Th2/Th1 ratios. Furthermore, we demonstrated that resistin directly skews the Th2/Th1 ratio via binding to monocytes. Thus, CMML granulocytes produce high levels of resistin, which contributes to defective monocyte maturation and induction of myeloid-derived suppressor cell features in monocytes, thereby fostering an immune-suppressive state in this disease.

The use of lipid nanoparticles (LNPs) to enhance delivery and immunogenicity of mRNA vaccines raises safety concerns due to strong reactogenicity. Developing naked mRNA vaccines, which can be delivered through liquid jet-injection to the skin, may offer a safer alternative, but their immunogenicity has been limited. To address this challenge, we screened clinically relevant adjuvants to enhance the efficacy of N1-methyl-pseudouridine-modified naked mRNA jet-injection vaccines in mouse models. Although many adjuvants showed little or inhibitory effects on inducing humoral or cellular immunity-despite their known ability to potentiate conventional non-mRNA vaccine platforms-we identified a combination of aluminum phosphate and CpG oligonucleotide that substantially enhanced both humoral and cellular immune responses with increased IgG2a/IgG1 ratio. This adjuvanted naked mRNA vaccine induced IgG2a and cellular immunity at levels comparable to those achieved with LNPs but with lower local and systemic reactogenicity. Mechanistic investigations revealed key roles for type I interferon and interleukin-6 signaling during immunization. Notably, this vaccine formulation in mice virtually eliminated infection by both SARS-CoV-2 and influenza A virus (A/California/7/2009 H1N1) in plaque-forming assays. Collectively, the adjuvanted naked mRNA jet-injection vaccine represents an effective, and safe platform against infectious diseases, with potential for clinic-ready translational applications.

Primary central nervous system (CNS) vasculitis is a rare and devastating inflammatory cerebrovascular disease for which pathogenic mechanisms remain poorly defined, largely due to limited access to human tissue and the absence of validated experimental models. Here, we present a mouse model of CNS-restricted vasculitis based on inducible, endothelial-biased neoantigen expression, termed brain endothelial ovalbumin (BrEndO). In BrEndO mice, tamoxifen-induced ovalbumin (OVA) expression in cerebral endothelial cells (ECs) elicited a controlled endogenous CD8+ T cell response at the blood-brain barrier that remained clinically restrained. In contrast, peripheral immune activation through infection with OVA-expressing Listeria monocytogenes or systemic PD-L1 immune checkpoint blockade converted this regulated response into severe, CD8+ T cell-driven CNS vasculitis with vascular injury and parenchymal infiltration. Transcriptomic profiling of inflamed cerebral microvessels and spatial transcriptomic analyses demonstrated shared inflammatory and immune activation signatures between BrEndO mice and human primary CNS vasculitis brain biopsies. These findings identify antigen-specific endothelial immune recognition as a regulated process whose breakdown precipitates CNS-restricted vasculitis and establish BrEndO as an experimentally tractable platform to investigate mechanisms and modifiers of immune-mediated cerebrovascular inflammation.

Heart failure (HF) is a leading cause of morbidity and mortality worldwide. Breathing instability and exaggerated chemoreflex activation contribute to adverse HF outcomes, yet the central mechanisms linking respiratory control to cardiac dysfunction remain poorly defined. The retrotrapezoid nucleus (RTN) is a key brainstem chemosensory region regulating ventilatory and autonomic output, but whether glial dysfunction within this circuit contributes to HF pathophysiology is unknown. We hypothesized that loss of astrocytic glutamate transporter 1 (GLT-1) in the RTN disrupts glutamate homeostasis, producing chemoreflex hyperactivity and cardiorespiratory dysfunction in HF. In a rat model of nonischemic HF, reduced astrocytic GLT-1 expression in the RTN was accompanied by extracellular glutamate accumulation and heightened excitability of chemosensitive neurons. These changes were associated with exaggerated central chemoreflex responses, breathing instability, autonomic imbalance, and impaired cardiac function. Astrocyte-specific GLT-1 restoration normalized RTN glutamate levels, reduced central chemoreceptor neuron hyperexcitability, stabilized breathing, and improved cardiac performance. Systemic riluzole, an FDA-approved glutamate-modulating drug, similarly reduced RTN glutamate accumulation and attenuated chemoreflex-driven respiratory instability. These findings identify loss of astrocytic GLT-1-mediated glutamate homeostasis in the RTN as a mechanism contributing to cardiorespiratory dysfunction in HF and support pharmacological targeting of brainstem glutamate homeostasis as a therapeutic strategy requiring further investigation.

..

Filamin C (FLNC) is an essential cytoskeletal protein required for cardiomyocyte structural integrity and mechanotransduction. Truncating FLNC variants cause haploinsufficiency associated with dilated and arrhythmogenic cardiomyopathies, yet no therapies currently address the underlying defect of FLNC deficiency. Gene replacement is challenging because the approximately 8.2-kb FLNC coding sequence exceeds the packaging capacity of conventional adeno-associated virus (AAV) vectors. Here, we engineered a triple-AAV strategy optimized for full-length FLNC reconstitution in the heart. Screening in HEK293 cells identified GP41-1 and IMPDH as a compatible orthogonal pair for efficient full-length FLNC reconstitution. In tamoxifen-inducible, cardiomyocyte-specific Flnc knockout mice, systemic delivery of optimized MyoAAV2A vectors restored full-length FLNC expression and re-established sarcomeric localization following cardiomyocyte FLNC depletion. A total vector dose of 1 × 10[13] genome copies/kg restored FLNC to approximately endogenous abundance, preserved systolic function and ventricular dimensions, limited cardiac enlargement, attenuated fibrotic remodeling and prevented premature death. FLNC replacement also attenuated pathological transcriptional remodeling and compensatory changes in cytoskeletal and junction-associated proteins. These findings demonstrate that systemic split-intein-mediated reconstitution of full-length FLNC can preserve cardiac structure and function in vivo, providing preclinical support for a mutation-independent therapeutic strategy for FLNC-associated cardiomyopathy.