N Nature Biomedical Engineering · Nov 17, 2025 HBsAg-tagged tumour vaccine system eliminates solid tumours through virus-specific memory T cells It is challenging for cancer vaccines to identify immunogenic antigens that are specifically and uniformly expressed on heterogeneous solid tumours and that can elicit production of T cells to lyse antigen-positive tumour cells and expand within the immunosuppressive tumour microenvironment. In contrast, microbial antigens are well-defined and robustly immunogenic and can activate specific memory T cells to eliminate microbes within the tumour microenvironment. Inspired by this, we developed a hepatitis B surface antigen (HBsAg)-tagged tumour vaccine system (H-TVAC). H-TVACleverages HBsAg-specific memory T cells from a HBsAg mRNA vaccine to target and lyse HBsAg-tagged tumour cells using the vaccinia virus. This approach also elicits a tumour-specific immune response through epitope spreading by recruiting dendritic cells, thereby eliminating heterogeneous solid tumours. In various preclinical murine models, including the B16-OVA, B16F10, MC38, CT26, 4T1 and H22 hepatocellular carcinoma, as well as a B16F10 bilateral tumour model, H-TVACdemonstrates anti-tumour immune responses, improved survival rates and reduced metastasis and recurrence. Cancer immunotherapy Tumour immunology biology mouse experiments
N Nature Biomedical Engineering · Oct 31, 2025 Steric stabilization-independent stealth cloak enables nanoreactors-mediated starvation therapy against refractory cancer The high interfacial energy of nanomaterials limits their certain biomedical applications that require stealthiness to minimize non-specific interaction with biological components. While steric repulsion-based entropic stabilization—such as PEGylation—has long been the dominant strategy for designing stealth nanomaterials, its inherent softness and susceptibility to dynamic deformation and external forces often result in only moderate stealth performance. Here we report a distinct approach to achieving stealthiness by harnessing an ion-pair network, rather than maximizing steric repulsion. Using model polyion complex nanoparticles composed of equimolar charge ratios of polycations and polyanions, we demonstrate that increasing crosslinks between the constituent polyions beyond a critical threshold effectively reduces protein adsorption and macrophage uptake, enabling prolonged circulation with a half-life exceeding 100 hours. Building on this, we develop an asparaginase-loaded vesicular nanoreactor enveloped by a semi-permeable ion-pair network sheath for asparagine starvation therapy. The extended circulation of these nanoreactors enables sustained depletion of asparagine, leading to improved therapeutic outcomes for metastatic breast and pancreatic cancers. Our findings open an avenue for improving the pharmacokinetics of nanomaterials for therapeutic delivery through delicately engineering stable intermolecular structures with holistic cooperativity. Cancer immunotherapy Cancer metabolism Drug delivery Metastasis Protein delivery biology
N Nature Biomedical Engineering · Oct 28, 2025 Computational design of synthetic receptors with programmable signalling activity for enhanced cancer T cell therapy The tumour microenvironment (TME) plays a key role in tumour progression, and soluble and cellular TME components can limit CAR-T cell function and persistence. Targeting soluble TME factors to enhance anti-tumour responses of engineered T cells through chimeric receptors is not broadly explored owing to the unpredictable signalling characteristics of synthetic protein receptors. Here we develop a computational protein design platform for the de novo bottom-up assembly of allosteric receptors with programmable input–output behaviours that respond to soluble TME factors with co-stimulation and cytokine signals in T cells, called TME-sensing switch receptor for enhanced response to tumours (T-SenSER). We develop two sets of T-SenSERs targeting vascular endothelial growth factor (VEGF) or colony-stimulating factor 1 (CSF1) that are both selectively enriched in a variety of tumours. Combination of CAR and T-SenSER in human T cells enhances anti-tumour responses in models of lung cancer and multiple myeloma, in a VEGF- or CSF1-dependent manner. Our study sets the stage for the accelerated development of synthetic biosensors with custom-built sensing and responses for basic and translational cell engineering applications. A computational method designs receptors called T-SenSERs, which have predictable signalling responses to tumour microenvironment soluble factors and can be co-expressed with conventional CARs in T cells to enhance their therapeutic activity. Cancer immunotherapy Protein design biology
N Nature Biomedical Engineering · Oct 23, 2025 Collagen-binding IL-12-armoured STEAP1 CAR-T cells reduce toxicity and treat prostate cancer in mouse models Immunosuppressive microenvironments, the lack of immune infiltration, and antigen heterogeneity pose challenges for chimaeric antigen receptor (CAR)-T cell therapies applied to solid tumours. Previously, CAR-T cells were armoured with immunostimulatory molecules, such as interleukin 12 (IL-12), to overcome this issue, but faced high toxicity. Here we show that collagen-binding domain-fused IL-12 (CBD-IL-12) secreted from CAR-T cells to target human six transmembrane epithelial antigen of prostate 1 (STEAP1) is retained within murine prostate tumours. This leads to high intratumoural interferon-γ levels, without hepatotoxicity and infiltration of T cells into non-target organs compared with unmodified IL-12. Both innate and adaptive immune compartments are activated and recognize diverse tumour antigens after CBD-IL-12-armoured CAR-T cell treatment. A combination of CBD-IL-12-armoured CAR-T cells and immune checkpoint inhibitors eradicated large tumours in an established prostate cancer mouse model. In addition, human CBD-IL-12-armoured CAR-T cells showed potent anti-tumour efficacy in a 22Rv1 xenograft while reducing circulating IL-12 levels compared with unmodified IL-12-armoured CAR-T cells. CBD fusion to potent payloads for CAR-T therapy may remove obstacles to their clinical translation towards elimination of solid tumours. Cancer immunotherapy Gene therapy Molecular medicine Protein delivery Synthetic biology biology mouse experiments
N Nature Biomedical Engineering · Oct 22, 2025 Targeting immunosuppressive myeloid cells via implant-mediated slow release of small molecules to prevent glioblastoma recurrence Glioblastoma is a highly aggressive brain tumour with a high risk of recurrence after surgery, even when combined with chemotherapy and radiotherapy. A major barrier to lasting treatment is the tumour’s immunosuppressive environment, which is largely dominated by myeloid cells. Here we describe the development of a biodegradable implant to sustainably release immune-modulator small molecules to reprogram tumour-infiltrating myeloid cells toward a pro-inflammatory, antitumour phenotype in the surgical cavity after tumour removal. In immunocompetent mouse models, this therapy induces interleukin-12 expression in myeloid cells without systemic cytokine elevation, and increases the infiltration of CD8+and CD4+T cells. Over 50% of mice treated (in combination with radiotherapy and chemotherapy) remain tumour-free during the experimental course (80 days). We further treated human glioblastoma explants ex vivo with the therapy and observed increased interleukin-12 expression in tumour-infiltrating myeloid cells, supporting the translational potential of this strategy. This implantable system offers a promising approach to prevent glioblastoma recurrence by activating innate immunity and sustaining immune surveillance post-surgery. Cancer immunotherapy Cancer microenvironment CNS cancer Drug delivery Immunotherapy biology mouse experiments
N Nature Biomedical Engineering · Oct 21, 2025 CCR5-targeted allogeneic gamma–delta CD19 chimeric antigen receptor T cells for HIV-associated B cell-malignancy immunotherapy Immune-based cell therapy offers a promising approach to cancer treatment. While autologous chimeric antigen receptor (CAR) T cells have shown success, production is time-consuming, costly and patient specific. Gamma–delta (γδ) T cells are promising for ‘off-the-shelf’ CAR T cell therapy. However, clinical translation of γδ CAR T cells is hampered by low frequency, resistance to genetic manipulation and advanced differentiation after expansion, limiting therapeutic feasibility. Here we demonstrate a method for in vitro activation and expansion of peripheral blood γδ T cells, facilitating high rates of gene editing and efficient CAR integration. Using artificial antigen-presenting cells, we produce minimally differentiated, highly functional γδ CAR T cells. By targeting a US Food and Drug Administration-approved CD19 CAR to the CCR5 locus, we generate CCR5-deficient γδ CD19 CAR T cells (γδ CCR5KI-CAR19), which demonstrated resistance to HIV-mediated depletion and robust antitumour responses against B cell lymphoma and leukaemia. γδ CCR5KI-CAR19 T cells enable the immunotherapy of HIV-associated B cell malignancies. These studies provide preclinical evidence supporting large-scale development of potent allogeneic γδ CAR T cells for diverse immunotherapies. Cancer Immunotherapy biology
N Nature Biomedical Engineering · Oct 01, 2025 Benchmarking foundation models as feature extractors for weakly supervised computational pathology Numerous pathology foundation models have been developed to extract clinically relevant information. There is currently limited literature independently evaluating these foundation models on external cohorts and clinically relevant tasks to uncover adjustments for future improvements. Here we benchmark 19 histopathology foundation models on 13 patient cohorts with 6,818 patients and 9,528 slides from lung, colorectal, gastric and breast cancers. The models were evaluated on weakly supervised tasks related to biomarkers, morphological properties and prognostic outcomes. We show that a vision-language foundation model, CONCH, yielded the highest overall performance when compared with vision-only foundation models, with Virchow2 as close second, although its superior performance was less pronounced in low-data scenarios and low-prevalence tasks. The experiments reveal that foundation models trained on distinct cohorts learn complementary features to predict the same label, and can be fused to outperform the current state of the art. An ensemble combining CONCH and Virchow2 predictions outperformed individual models in 55% of tasks, leveraging their complementary strengths in classification scenarios. Moreover, our findings suggest that data diversity outweighs data volume for foundation models. Cancer imaging Tumour biomarkers Cancer Machine Learning Human Clinical
