An in-situ vaccine for cancer immunotherapy can empower the anti-tumor immune response in mice and can shrink tumors when combined with other immunotherapy strategies, according to new research. Although early, the results suggest that the vaccine platform should be evaluated further as another addition to the immunotherapy toolbox for solid tumors. Checkpoint blockade is the most well-known form of immunotherapy, but there are other strategies to enhance the immune system’s ability to wipe out tumors. One approach involves encouraging immunogenic cell death, which occurs when cells release certain stress-related signals that prod immune cells to gobble them up. This allows researchers to promote immune responses by exploiting antigens already available in tumors instead of relying on previously isolated antigens, in effect creating an in-situ cancer vaccine. However, more work is needed to streamline immunogenic cell death strategies and to overcome obstacles in drug delivery. Here, Dingkang Liu and colleagues generated and tested a modular in-situ vaccine platform which uses the cage-like protein ferritin as a framework and as a delivery vehicle for the peptide KLA. However, their platform delivers the treatment via a prodrug that only activates after being exposed to the enzyme MMP14 within tumors. This circumvents issues with drug leakage and interception by the liver that have held back previous ferritin-based strategies. When tested in mice, the platform delivered its therapeutic payload into breast cancer tumors and successfully elicited potent in-situ immune responses. Furthermore, it potentiated the tumor-shrinking effects of anti-PD-1 checkpoint blockade, leading Liu et al. to speculate that the platform could find broad applicability in overcoming immune tolerance and evasion by tumors.

The Context-Guided Segmentation Network (CGS-Net) developed by University of Maine researchers introduces a deep learning architecture designed to interpret microscopic images of tissue with greater precision than conventional AI models. Powered by a dual-encoder model that mirrors the workflow of a pathologist examining a slide, one branch of the network processes a high-resolution image patch to capture cell-level details, while the other examines a lower-resolution patch encompassing the surrounding tissue. A system of interconnected encoders and decoders uses data from both the high and low resolution images for a complete analysis. 

When under cellular stress, breast cancer cells turn on genes that promote tumor growth and stress resistance.

UC San Diego researchers have developed a new targeted therapy for triple-negative breast cancer, the most difficult-to-treat subtype of the disease.