Letting Bacteria Infiltrate Oxygen-Poor Tumors: Engineered Probiotic Delivers IL-2 Inside Pancreatic Tumors in Mice

Pancreatic cancer not only makes it difficult for drugs to penetrate deeply, but is also adept at keeping immune attacks out. A University of Chicago team is now trying to link these two weaknesses: using engineered bacteria that favor low-oxygen environments to enter tumors and produce, on site, a signaling molecule capable of awakening anticancer immune cells. This living therapy, called BifidoSumIL-2, suppressed pancreatic tumors in mice and improved the effects of chemotherapy, radiotherapy, and immunotherapy.

The researchers engineered Bifidobacterium longum to secrete SumIL-2, a modified interleukin-2 variant. Conventional IL-2 can promote the proliferation of T cells and natural killer cells, but it may also activate regulatory T cells that suppress immune responses; high-dose systemic use can also cause severe toxicity. SumIL-2 was designed to bias signaling toward effector T cells rather than regulatory T cells.

The bacteria serve as a platform for delivery and local production. Bifidobacterium longum is anaerobic and, after systemic injection, can accumulate in oxygen-poor regions of tumors. Healthy tissues with more abundant oxygen are less conducive to its growth, potentially narrowing the area exposed to IL-2. However, this tumor selectivity is currently supported only by animal experiments and cannot yet be equated with precise targeting in humans.

The study used both subcutaneous and orthotopic mouse models of pancreatic ductal adenocarcinoma; in the latter, tumors were established within the pancreas, better preserving the relevant anatomical and immune environment. After BifidoSumIL-2 treatment, the ratio of effector to regulatory T cells within tumors shifted in an antitumor direction, the activity of CD8-positive T cells and natural killer cells increased, immunosuppressive myeloid cells decreased, and tumor growth slowed.

A more consequential test of this strategy is whether it can loosen pancreatic cancer’s immune barrier for existing therapies. In the study, BifidoSumIL-2 was paired separately with gemcitabine, radiotherapy, and an anti-PD-L1 checkpoint inhibitor. Tumor control and mouse survival were both better than with the corresponding monotherapies. The results support the possibility that different treatment mechanisms can complement one another, but they are not sufficient to determine which combination is best suited for clinical use, nor can benefits for patients be inferred directly from mouse survival.

Living bacterial drugs also raise issues distinct from those associated with conventional protein drugs before they can be used in humans: where the bacteria remain, whether they can be reliably cleared, whether long-term or repeated administration causes infection or unintended immune responses, and how long efficacy can be maintained all require further study. The team also proposed the possibility of oral delivery, but this study used systemic injection, and the two routes cannot be considered to have demonstrated equivalent effects.

These findings remain preclinical evidence, with no human safety or efficacy data. The University of Chicago has filed a provisional patent application for the related technology. In addition to expanded assessments of toxicology, strain stability, and biosafety, the next steps must include establishing manufacturing and regulatory frameworks capable of controlling the quality and in vivo behavior of living bacterial products.