Cancer’s deadliest move is often not the growth of the original tumor, but its ability to travel. Metastasis, the spread of malignant cells from a primary site to distant organs, is responsible for at least two-thirds of cancer deaths. Yet many experimental drugs designed to block this process fail when they reach clinical trials. One reason is that researchers still lack reliable laboratory models that reproduce the complex interactions between human cancer cells, blood vessels and the tissues they invade. Now, engineers and cancer biologists at Columbia University have developed a human “multi-organ chip” designed to model how breast cancer cells leave the bloodstream and colonize bone and lung tissue. The platform combines engineered human tissues, a vascular channel and controlled fluid flow in a single system, offering a detailed view of metastatic progression that conventional laboratory models cannot provide.
The study, published in Science Translational Medicine, focuses on organ colonization, one of the least understood stages of metastasis. Cancer cells first detach from a primary tumor, enter the circulation and survive the physical and biological challenges of traveling through the bloodstream. They must then attach to the inner surface of blood vessels, cross the endothelial barrier that separates blood from tissue, and adapt to a new cellular environment. Only a small fraction of circulating tumor cells complete this journey successfully, but those that do can establish secondary tumors. The Columbia team sought to recreate this sequence in a controllable human tissue model, allowing researchers to observe not only where cancer cells travel, but also how they alter the tissues they encounter.
The chip was developed by a team led by Gordana Vunjak-Novakovic, a University Professor at Columbia University and a professor of biomedical engineering and medical sciences. It contains separate compartments holding millimeter-scale engineered human bone and lung tissues, connected through a vascular circulation system. The tissues were created from induced pluripotent stem cells, or iPSCs, which can be reprogrammed into different human cell types. Tissue-specific scaffold and bioreactor systems were used to guide the cells toward mature bone and lung structures. The design allowed each tissue to be maintained under conditions suited to its own biology while remaining connected to the same circulating fluid. This arrangement enabled the researchers to study communication between distant organs without relying on an entire animal.
A central feature of the device is its engineered endothelial barrier. Endothelial cells form the thin inner lining of blood vessels and regulate the movement of cells and molecules between the bloodstream and surrounding tissues. In metastasis, this barrier can become a decisive checkpoint: cancer cells must adhere to the endothelium, trigger or exploit changes in its permeability, and force their way into the underlying tissue. On the chip, the endothelial layer separates the vascular channel from the bone and lung compartments while remaining selectively permeable. This means that nutrients, signaling molecules and other soluble factors can circulate between compartments, while the movement of cells can be monitored more precisely. Researchers can therefore examine the physical process of tumor-cell extravasation, the term used for passage from blood into tissue, under human-relevant conditions.
To test the system, the scientists introduced human breast cancer cells into the vascular circulation. They compared cancer cell populations known to display different patterns of organ preference, sometimes described as tissue tropism. Cells with a tendency to spread to bone showed stronger colonization of the engineered bone compartment and produced more pronounced signs of bone degeneration. Cells associated with lung-directed metastasis caused greater disruption within the lung tissue and showed only modest colonization of bone. These contrasting behaviors suggest that the chip captured important organ-specific features of metastasis rather than simply allowing cancer cells to attach randomly to any available tissue. The observed patterns were accompanied by differences in the factors secreted by the cancer cells and the tissues, providing molecular clues about how metastatic preferences are established.
The model also revealed evidence of pre-metastatic niche formation, a process in which a primary tumor begins preparing distant organs before cancer cells arrive. Tumors can release proteins, nucleic acids and extracellular vesicles into the circulation, prompting changes in remote tissues. These signals may recruit immune cells, remodel the extracellular matrix or alter blood-vessel behavior, making an organ more receptive to incoming tumor cells. In the Columbia experiments, post-analysis of both engineered tissue compartments showed signs that the cancer cells were conditioning the bone and lung environments. This finding is significant because the pre-metastatic niche is difficult to isolate and study in patients, where many biological processes occur simultaneously. A connected human chip makes it possible to track these early changes in a controlled setting and determine which signals help distant tissues become vulnerable.
The platform’s human origin is particularly important for drug development. Animal models have provided essential insights into tumor biology, but rodents and humans differ in immune responses, tissue architecture, metabolism and the molecular signals that govern cancer spread. Such differences can affect whether a treatment appears effective before entering clinical testing. The new system is not intended to replace every animal experiment, nor does it reproduce the full complexity of a human body. However, it can complement animal studies by offering a human-cell environment in which researchers can test specific mechanisms of metastasis. Because the tissues can potentially be generated from patient-derived or genetically defined cells, the approach may eventually support personalized studies of why one patient’s cancer spreads preferentially to bone while another patient’s disease targets the lungs.
The researchers say the chip could also be used to investigate therapeutic targets that are difficult to evaluate with standard two-dimensional cell cultures. In a conventional dish, cancer cells often grow on flat plastic surfaces and interact with a limited set of cell types. They do not experience realistic fluid shear, three-dimensional extracellular matrix, endothelial barriers or the biochemical exchange between multiple organs. The multi-organ chip incorporates these features into a microscale system in which flow, cell composition and tissue exposure can be controlled. Investigators could use the device to test drugs that block tumor-cell adhesion, prevent endothelial crossing, disrupt pre-metastatic signaling or preserve bone and lung tissue after cancer-cell arrival. Repeated measurements of secreted molecules, tissue damage and cell behavior could help identify why a treatment works in one organ but fails in another.
The study reflects a broader shift toward microphysiological systems, also known as organs-on-chips, as regulators and funding agencies place greater emphasis on human-based methods for biomedical research. Columbia researchers Ilaria Baldassarri, one of the study’s lead authors, Andrea Califano, Peter Sims and Hanina Hibshoosh contributed to the work alongside Vunjak-Novakovic’s tissue-engineering group. Their findings demonstrate how engineered tissues can be linked to study a disease that is inherently systemic: cancer cells move through the circulation, communicate with distant organs and reshape the environments they enter. By providing access to these interactions in real time, the system may help researchers uncover molecular pathways that remain hidden in isolated cultures. Its long-term value will depend on further validation with additional cancer types, patient samples and therapeutic studies, but the model offers a promising new route for exploring metastasis as a dynamic conversation between tumor cells, blood vessels and organ-specific tissues.
Subject of Research: Human multi-organ chip modeling breast cancer metastasis to bone and lung
Article Title: Organ-specific colonization and niche remodeling in a human tissue model of metastasis
News Publication Date: 19-Aug-2026
Web References: https://www.engineering.columbia.edu/faculty-staff/directory/gordana-vunjak-novakovic ; https://www.bme.columbia.edu/ ; https://gvnlab.bme.columbia.edu/
References: Science Translational Medicine, “Organ-specific colonization and niche remodeling in a human tissue model of metastasis”
Image Credits: Steve Zill/Columbia Engineering
Keywords: Cancer, metastasis, breast cancer, organ-on-a-chip, multi-organ chip, tissue engineering, bone metastasis, lung metastasis, induced pluripotent stem cells, microphysiological systems, pre-metastatic niche, cancer research
Tags: advancements in organ-on-chip technology for cancerblood vessel and tissue interaction in metastasisblood-brain barrier and cancer cell migrationbreast cancer cell disseminationcancer metastasis modelingengineering human tissues for cancer studieshuman multi-organ chip for cancer researchin vitro models of cancer metastasislimitations of traditional metastasis modelsorgan-specific colonization in cancer spreadstudying cancer cell detachment and invasionvascular channel design in cancer chips
