KRAS inhibitors (KRASi) are reshaping the treatment of aggressive tumors such as pancreatic ductal adenocarcinoma (PDAC), but it is appreciated that patients will eventually acquire resistance through multiple pathways. To avoid this outcome, there is much effort to identify combination therapies that take advantage of the 'window-of-opportunity' permitted from KRAS inhibition. Tumor cell-intrinsic targeting - as a result of KRASi - results in altered immune and stromal features within the PDAC tumor site that could allow for improved, non-KRAS-targeted, treatment efficacy. Qiang and colleagues investigate the potential of pairing KRASi with immunotherapy - here, an IL-21 mimic - for therapeutic benefit during this 'open window' phase of the KRASi-induced tumor debulking. IL-21 promotes a CD4 T cell response that requires both conventional type 2 dendritic cells (cDC2s) and interferon (IFN)-γ, but not direct tumor cell recognition via major histocompatibility complex (MHC). This work fits in a series of studies highlighting the ability to drive PDAC tumor rejection through cytokines derived from CD4, but not CD8, T cells. These findings highlight a divergence from more 'traditional' immunotherapeutic responses, where cytotoxic CD8 T cells have historically dominated tumor control. Identifying alternative paths towards targeting PDAC may be beneficial - and points towards cell types and modalities that could be leveraged for clinical benefit as the field identifies potential combinatorial partners for KRASi and more patients start receiving these drugs.

Vasculogenic mimicry (VM) is a process in which malignant cells acquire endothelial-like functions to generate perfusable vascular channels independent of endothelial cells. VM represents a major barrier to effective vascular endothelial growth factor A (VEGFA)-targeted therapy in cancer. In this study, we used organ-scale vascular imaging that identified a mechanism of VM in glioblastoma (GBM) mediated by alpha-L-fucosidase 1 (FUCA1). FUCA1 was preferentially expressed in glioblastoma stem-like cells (GSCs) with high CD44 expression. FUCA1 functioned as a key regulator of GSC-mediated VM by stabilizing LDHA protein and promoting a metabolic shift toward glycolysis. Elevated glycolytic flux increased lactate production and enhanced histone H3 lysine 18 lactylation (H3K18la), which in turn activated SERPINF1 transcription and drove VM in GSCs. In preclinical models, the FUCA1 inhibitor deoxyfuconojirimycin hydrochloride (DFH) significantly reduced GSC tumorigenicity and VM in vivo and exhibited robust antitumor activity when combined with anti-VEGFA therapy. Collectively, these findings define a FUCA1-mediated histone lactylation pathway that enhances GSC VM and highlight a therapeutically targetable mechanism to improve the efficacy of current anti-angiogenic strategies in GBM.

Metabolic reprogramming is a well-described hallmark of clear cell renal cell carcinoma (ccRCC). Gene copy number variation and altered gene expression data suggest enhanced branched-chain amino acid (BCAA) transamination and reduced BCAA oxidation are highly selected traits in ccRCC. Surprisingly, we found that genetic modulation of BCAA oxidation had little effect on experimental ccRCC cell growth in vitro and in vivo. In contrast, BCAA transamination in ccRCC cells was essential for the generation of glutamate-derived glutathione and cell viability when glutamine was limiting. Thus, BCAA transamination conferred resistance to states of low glutamine or treatment with the glutaminase inhibitor CB-839 by providing glutathione independent of glutamine breakdown. Targeting BCAA transamination in vitro and in vivo sensitized ccRCC to CB-839 mediated glutaminase inhibition by decreasing glutathione biosynthesis and redox homeostasis. These findings reveal that BCAA transamination maintains critical glutamate and glutathione levels in glutamine-deprived ccRCC, explaining why CB-839 treatment has been ineffective in ccRCC clinical trials and demonstrate an approach to treat ccRCC by targeting BCAA transamination to create a vulnerability to glutaminase inhibition.

Oncolytic virotherapy is a promising cancer treatment that elicits antitumor immunity. However, the efficacy of virotherapy is often constrained by tumor-intrinsic immune suppression. To identify conserved virus-induced programs driving immune escape, we applied multi-omics and functional assays across multiple tumor and virus models, uncovering an SLC3A2-driven immunosuppressive metabolic barrier. In oncolytic herpes simplex virus (oHSV)-infected tumors, a biphasic immune trajectory emerged: early CD8+ T cell activation was followed by SLC3A2-mediated methionine and tryptophan uptake, driving T cell exhaustion. Mechanistically, oHSV upregulated SLC3A2 via the m6A-IGF2BP3 axis, while tumor-intrinsic tryptophan metabolism engaged the AHR/ARNT/ICP4 cascade to amplify viral replication and reinforce the immunosuppressive program. Importantly, dietary methionine supplementation disrupted the metabolic checkpoint, restored T cell effector function, and synergized with oHSV expressing anti-PD1 to enhance tumor control in humanized preclinical models. Notably, a parallel SLC3A2-centered metabolic-immunologic circuit operated in vesicular stomatitis virus (VSV) infection. Together, these findings delineate an actionable, conserved metabolic checkpoint and provide a framework for integrating metabolic modulation into cancer viroimmunotherapy.

Barrett's esophagus (BE) patients require improved cancer risk stratification to identify high-risk individuals while reducing over-treatment in low-risk populations. BE evolution towards adenocarcinoma is likely driven by a combination of genomic and microenvironmental factors, suggesting the potential of predictive models that integrate both. Here, we developed BEACON (Barrett's Esophagus DNA content Abnormality and immune ecology for Cancer Outcome), a spatially aware framework integrating deep learning-based prediction of DNA content and immune architecture from routine histopathology. First, DACOR, a multi-instance learning model that predicts DNA content from histopathology, was trained using 777 BE biopsies with flow cytometry-based DNA content data scanned at two institutions, and it demonstrated strong performance across both validation and test datasets for DNA content abnormality prediction. Patients classified as DNA content abnormal by DACOR demonstrated increased cancer progression. Complementary models for cell classification and tissue segmentation enabled spatial immune ecology metric computations. Abnormal regions exhibited increased lymphoplasmacytic inflammation versus normal regions. Among patients with DNA content abnormality, cancer progressors exhibited increased plasma cell clustering adjacent to abnormal epithelium compared to non-progressors. Finally, a logistic regression model was developed that integrated molecular, immune ecological, and epithelial morphology features for cancer risk stratification. The integrated risk classification model stratified DNA content abnormal patients into high- and low-risk groups that corresponded with progression and patient survival. Overall, BEACON spatially integrates molecular abnormality with immune ecology to stratify BE patients by cancer progression risk using routine pathology images.

Hormone receptor positive (HR+) breast cancer is the most common breast cancer subtype globally and remains a leading cause of cancer-related death in women. Black women diagnosed HR+ breast cancer experience significantly worse outcomes than White women, a disparity that persists even after accounting for factors such as healthcare access and educational attainment. Understanding the biological mechanisms underlying these differences has been limited by the underrepresentation of Black patients in genomic datasets. To address this gap, we generated an HR+ breast cancer transcriptomic dataset enriched for Black patients while controlling for access to healthcare and education level. Tumors from Black women showed enrichment of tumor microenvironment-related signatures, including increased collagen deposition and prognostically unfavorable T-cell landscapes. Data from in vitro and in vivo experimental models showed that race-aligned collagen deposition patterns were driven, at least in part, by tumor cell-intrinsic signaling and promoted metastatic phenotypes associated with poorer outcomes. Further analysis of the patient transcriptomic dataset revealed that unfavorable T-cell signatures previously attributed to race or ancestry were more strongly associated with poverty burden. Across multiple independent datasets, STAT4 emerged as a potential master regulator of this poverty-associated immune phenotype. Collectively, these findings suggest that molecular features of breast tumors are shaped by both biological and structural factors linked to self-identified race and socioeconomic disadvantage. These results demonstrate that integrating social determinants of health with molecular analyses can reveal mechanisms contributing to cancer outcome disparities and provide a foundation for the development of more effective and equitable precision oncology approaches.

Gastric mixed neuroendocrine-non-neuroendocrine neoplasms (MiNENs) are aggressive malignancies that exhibit complex cellular heterogeneity driven by the dual lineages. The dismal patient survival rates highlight the need to elucidate the cellular origin of gastric MiNENs and the key regulatory mechanisms to develop improved treatment strategies. Here, we identified a gastric POU2F3+epithelial population as a potential cell of origin for gastric MiNENs. Lineage tracing and single-cell RNA sequencing further identified that POU2F3+ cells include three distinct subpopulations with divergent differentiation trajectories: progenitor cells (SOX2+SOX9+), tuft cells (POU2F3+DCLK1+), and neuroendocrine cells (SYP+CHGA+). Combined Rb1/Trp53 deletion and MYC overexpression in POU2F3+ epithelial cells (Pou2f3-CreERT; Rb1fl/fl; Trp53fl/fl; LSL-MycT58A) led to the rapid development of gastric MiNENs containing both neuroendocrine and non-neuroendocrine lesions, recapitulating human gastric MiNENs pathology. Transcription factor activity analysis and spatial transcriptomics revealed SOX2 and SOX9 as critical regulators driving subpopulation specification, with SOX2 and SOX9 being enriched in dysplasia and adenocarcinoma, respectively. Bulk RNA sequencing further showed increased expression of CXCL5 in gastric MiNENs, and inhibition of CXCL5 reduced the growth of tumor organoids and xenografts in mouse models. Together these findings offer a physiologically relevant model for studying gastric MiNENs and identify CXCL5 as a potential therapeutic target for this heterogeneous cancer.

Mutations in isocitrate dehydrogenase (IDHm) define a distinct molecular class of gliomas. IDHm converts α-ketoglutarate (α-KG) to the oncometabolite D-2-hydroxyglutarate (D-2HG), which drives tumorigenesis. The IDHm inhibitor vorasidenib suppresses D-2HG production and prolongs progression-free survival in some, but not all, patients. Here, using patient-derived models, genetically engineered glioma cells, orthotopic xenografts, and patient tissue, we showed that phosphoglycerate dehydrogenase (PHGDH) mediates intrinsic resistance to vorasidenib by sustaining D-2HG production despite IDHm inhibition. Silencing PHGDH suppressed residual D-2HG synthesis and restored vorasidenib sensitivity in resistant models, whereas PHGDH overexpression restored D-2HG production and induced resistance in sensitive models. Importantly, restoration of intracellular D-2HG rescued growth despite continued vorasidenib treatment, demonstrating a causal role for residual D-2HG production in the resistant phenotype. Mechanistically, a permissive chromatin state characterized by increased H3K4me3, reduced H3K27me3, and enhanced ATF4 occupancy at the PHGDH promoter facilitated PHGDH upregulation in resistant models. Pharmacologic induction of this chromatin state induced PHGDH expression, restored D-2HG production, and conferred resistance to vorasidenib in sensitive models. Consistently, tumor tissue from patients with poor response to IDHm inhibitor treatment exhibited higher PHGDH expression, coupled with the ATF4-enriched permissive chromatin state observed in resistant models. Finally, deuterium metabolic imaging of D-2HG production from diethyl-[3,3'-2H]-α-KG provided an early, non-invasive readout of response and resistance to vorasidenib in vivo. Collectively, this study identifies a chromatin-regulated ATF4-PHGDH-D-2HG axis as a mechanism of intrinsic resistance to IDHm inhibition and establishes diethyl-[3,3'-2H]-α-KG as a metabolic imaging probe for interrogating treatment response in IDHm gliomas.

Pancreatic ductal adenocarcinoma (PDAC) is a lethal malignancy 20% more likely to develop in Black African Americans (BAA) compared to Non-African Americans (NAA). Large-scale gene expression analysis of tumor subtypes and cellular microenvironments in a diverse PDAC cohort could help provide insights into cell and molecular differences in BAA cancer cells. Here, we conducted a multi-omics interrogation of BAA versus NAA PDAC using whole exome sequencing, spatial transcriptomics (ST), and spatial proteomics. Whole exome sequencing confirmed patient ancestry and showed no significant differences in driver mutations between race groups. BAA PDAC tumors had significantly fewer classical molecular subtype cells with no change in more aggressive basal-like cells. Instead, BAA patients showed a significant enrichment of the intermediate/hybrid subtype compared to NAA patients. BAA tumors also displayed increased B cell and reduced myeloid cell interactions compared to NAA tumors. Analysis of tertiary lymphoid structures in a validation cohort revealed that tumors in BAA patients displayed larger TLS size and number compared to NAA. In conclusion, these findings identify an increased intermediate molecular subtype signature in BAA tumors at the expense of cells with a classical subtype, accompanied by a distinct immune cell composition and spatial distribution. This study highlights that health disparities in patients of different ancestries may have a molecular basis that influences tumor aggressiveness and the immune response, which may inform therapeutic strategies.

Growth-induced solid stress and fluid pressure contribute to tumor progression and drug resistance. Beyond tumor endogenous physical forces, cells in the tumor microenvironment may also experience external macroscopic mechanical stress, and some tumors preferentially develop in weight-bearing regions with higher focal pressure. Characterization of the impact of macroscopic mechanical stress on the tumor microenvironment could help uncover the mechanistic links between external mechanical forces and increased tumor incidence. Here, we showed that recurrent macroscopic mechanical stress disrupts nuclear envelope integrity, leading to spatial reorganization and fibroblast subtype changes at the tumor-stromal interface. In mouse models of plantar melanoma, exogenous mechanical stress induced nuclear envelope rupture in melanoma cells and increased LRRC15-positive myofibroblasts. Reduction of nuclear envelope integrity through inducible deletion of Lmna likewise increased LRRC15-positive myofibroblast generation. Co-culture with myofibroblasts promoted nuclear envelope rupture in melanoma cells in 3D culture, suggesting a positive feedback loop. In addition, mechanical stress-driven transcriptomic changes in tumor and stromal cells were observed alongside enhanced lymph node metastasis and immunotherapy resistance. Analysis of human melanoma samples revealed mechanical stress-associated stromal remodeling and a higher incidence of lymph node metastasis in tumors arising in mechanically stressed regions, although these features were not significantly associated with overall survival. Together, these findings suggest that macroscopic mechanical stress may influence cancer progression by inducing nuclear envelope damage and reshaping tumor-stromal interactions.

Targeted drug conjugates often rely on receptor-mediated internalization and intracellular trafficking for payload release. This requirement limits the range of tumor-associated targets that can be effectively exploited and has motivated the development of strategies for extracellular drug release. Wen and colleagues introduce a binding-to-release (BTR) strategy in which target engagement itself serves as the trigger for payload release. The authors used ligand-directed chemistry to develop phosphorus(V)-phenol exchange (PhoPEx) chemistry that exploits proximity between an electrophilic linker and a nucleophilic tyrosine residue within the binding pocket of the target to initiate target-triggered payload release. Through optimization of the phenol-based scaffold, the authors evaluated 15 PhoPEx compounds to balance hydrolytic stability and reaction kinetics. The authors then used fibroblast activation protein, a poorly internalizing stromal target, to test the BTR strategy using a membrane-permeable payload, resulting in improved tumor drug exposure and antitumor activity in preclinical models. The BTR strategy was further extended to imaging applications and other targets, including programmed death-ligand 1. Although demonstrated primarily with small-molecule conjugates, BTR introduces an alternative approach for targeted drug delivery of poorly internalizing targets in which target binding could be exploited as a trigger, potentially expanding the landscape of tumor-associated targets accessible to targeted therapeutics.

The exquisite antigen specificity of T cells is one foundation of modern cancer immunotherapy. The therapeutic efficacy of checkpoint blockade, cancer vaccines, and engineered T-cell therapies, all appear to depend on T cells capable of recognizing specific antigens displayed by tumor cells. In a recent issue of Nature Immunology, Smith and colleagues challenge this paradigm by demonstrating that local activation of virus-specific memory T cells can eliminate tumors without direct tumor-cell recognition. Instead, activated T cells initiate a multicellular inflammatory response involving IFNγ, TNF, nitric oxide and recruited myeloid cells that collectively induce tumor-cell death. These findings separate the antigen specificity required for T-cell activation from that required for tumor killing and expand the role of T cells from direct cytotoxic effectors to organizers of a broader antitumor response. Together, these findings broaden the potential scope of T-cell-mediated antitumor immunity beyond the constraints of direct tumor-antigen recognition.

The RAS/MAPK pathway is among the most frequently dysregulated in human cancer, predominantly driven by activating mutations across RAS and BRAF in a broad spectrum of tumor types. The recent clinical validation of direct inhibitors of RAS, long considered undruggable, has reenergized the field and renewed focus on targeting the full RAS/MAPK axis. MEK, a central node in MAPK signaling, remains an attractive mutation-agnostic target, yet the efficacy of first-generation MEK inhibitors has been limited by pathway reactivation in tumors and by on-target, off-tissue toxicity. In this issue of Cancer Research, Kolitz and colleagues describe atebimetinib, a novel MEK inhibitor that also prevents feedback-induced MEK phosphorylation by RAF. The uniquely short plasma half-life of atebimetinib enables deep cyclic inhibition (DCI): transient but profound MEK/ERK output suppression with full recovery between doses. At each drug's tolerability limit, atebimetinib outperformed the MEK inhibitor binimetinib across multiple RAS- and BRAF-mutant tumor models. Whether DCI or suppression of MEK phosphorylation is the primary driver of the improved therapeutic window and how atebimetinib compares with the clinical standard and other MEK-targeting agents remain important open questions, but the work makes a compelling case that novel pharmacology can unlock therapeutic spaces at targets long considered exhausted. See related article by Kolitz et al., p. 5223.

Lung squamous cell carcinoma (LUSC) is an aggressive subtype of non-small cell lung cancer with limited targeted therapeutic options and incompletely defined immune-regulatory dependencies. Here, we identified atypical chemokine receptor 3 (ACKR3) as a squamous-enriched regulator that restrains malignant progression in LUSC. Integrative analyses of human LUSC datasets, autochthonous genetically engineered mouse models, and ex vivo organoid systems showed that ACKR3 loss accelerated tumor growth and remodeled the immune microenvironment, marked by reduced CD8[+] T cell infiltration and increased T cell exhaustion. Mechanistically, ACKR3 limited activation of the MIF-CD74-PD-L1 axis, thereby constraining tumor-cell PD-L1 expression and preserving antitumor immune surveillance. In preclinical LUSC models, treatment with the ACKR3-directed PD-L1 degrader KineTAC, a bispecific molecule that co-engages ACKR3 and PD-L1 to promote lysosomal degradation of PD-L1, enhanced antitumor efficacy, particularly in combination with chemotherapy. Together, these findings define ACKR3 as a squamous-associated immune checkpoint regulator and nominate ACKR3-directed PD-L1 degradation as a tumor-selective therapeutic strategy for LUSC.

Hepatocellular carcinoma (HCC) driven by gain-of-function mutations in CTNNB1 (CTNNB1GOF) that stabilize β-catenin exhibits a profoundly immunosuppressive tumor microenvironment (TME), rendering it resistant to immune checkpoint blockade (ICB). Metabolic reprogramming is increasingly recognized as a crucial mediator of immune activity, highlighting the need to elucidate the specific metabolic mechanisms underlying immune exclusion in CTNNB1GOF HCC. Here, we identified a critical metabolic-immune network orchestrated by co-mutations of CTNNB1 and ARID2. Integrative analysis of HCC patient cohorts and a de novo MYC/CTNNB1GOF HCC mouse model revealed that tumors harboring concurrent CTNNB1 and ARID2 mutations exhibited a more immune-active TME and improved survival compared to tumors with CTNNB1 mutations alone. Mechanistically, β-catenin and ARID2 cooperatively enhanced the transcription of tryptophan 2,3-dioxygenase (TDO2), leading to accelerated kynurenine production and activation of the aryl hydrocarbon receptor (AhR) in myeloid-derived suppressor cells (MDSCs). This metabolic axis established an immunosuppressive milieu characterized by MDSC accumulation and impaired T cell function. Genetic ablation of Arid2 in CTNNB1GOF HCC attenuated TDO2-dependent kynurenine synthesis, reduced MDSC infiltration, and enhanced CD8[+] T cell cytotoxicity, thereby sensitizing tumors to anti-PD-1 therapy. Clinically, co-mutations of CTNNB1 and ARID2 correlated with elevated immune infiltration and favorable ICB responses across multiple human cancers, including HCC. Importantly, pharmacological inhibition of AhR phenocopied the immunomodulatory effect of ARID2 loss and, when combined with anti-PD-1, robustly reversed the immune-cold phenotype of CTNNB1GOF HCC. These findings reveal TDO2-kynurenine-AhR signaling as a pivotal metabolic pathway enforcing immune evasion in CTNNB1-mutant HCC.