Esophageal adenocarcinoma (EAC), one of two major forms of esophageal cancer, is the sixth most deadly cancer worldwide for which no effective targeted therapy exists. Patients need to rely on chemotherapy as a standard-of-care, which is started ahead of surgical interventions as a so-called “neoadjuvant chemotherapy” (NACT) in the hope of shrinking or controlling tumors. However, most patients become resistant to certain NACTs, leading to poor outcomes.

Given the utter lack of therapeutic alternatives, responders and non-responders alike continue to receive one of the available chemotherapies without knowing whether it will work. Even in responders, the chemotherapy of choice may not completely stop their tumors from progressing and metastasizing, and it can have toxic side effects on the body. The availability of a personalized, patient-specific precision oncology model that can accurately predict a patient’s response to different NACTs in a timely manner is a critical unmet need.

Researchers had grown so-called “organoids” from biopsied EAC cells, which are 3D esophageal mini-organs formed with tissue-specific stem cells that exhibit critical features of the esophageal epithelial lining. However, these lack important components of a patient’s specific tumor microenvironment (TME), such as the stromal fibroblasts and collagen fibers, and thus, they do not show the same responses to NACT as actual tumors.

Now, a research collaboration led by Donald Ingber, M.D., Ph.D., Founding Director at the Wyss Institute for Biologically Inspired Engineering at Harvard University and Lorenzo Ferri, M.D., who heads the Division of Thoracic and Upper Gastrointestinal Surgery at the McGill University Health Centre in Montreal, has advanced a personalized medicine solution with the potential to improve chemotherapy for EAC patients.

The researchers leveraged the Wyss Institute’s human Organ Chip microfluidic culture technology and used it to co-culture EAC organoids next to stromal cells isolated from the same biopsies that the McGill team obtained from EAC patients in a clinical cohort study to create patient-specific, TME-inclusive Cancer Chip models. By recapitulating some of the inherent TME complexity in vitro, the team was able to predict patients’ tumor responses to the standard NACT much more accurately than more static, less complex 3D organoid models.

Since the approach can produce results within 12 days from starting the model, it enables the rapid stratification of EAC patients into responders and non-responders, and investigation of non-standard NACTs based on different chemotherapy agents for resistant patients in a clinically useful timeframe. The findings are reported in the Journal of Translational Medicine.

“This patient-centered approach strongly builds on our previous successes using human Organ Chip technology to recapitulate each individual cancer patient’s TME outside their body so that we can identify the drug combination that will work best for that very patient. This new way to approach personalized medicine could be implemented at clinical centers focusing on the care of patients suffering from many different types of cancer, such as the one run by our collaborators with patients who have esophageal cancer,” said Ingber.

“Perhaps equally important, it can also be used as a pre-clinical testbed to break new ground in the development of tumor- or stroma-targeted therapies for cancer patients and enable the discovery of biomarkers that could be used to monitor and optimize drug effects in these patients.”

Ingber is also the Judah Folkman Professor of Vascular Biology at Harvard Medical School and Boston Children’s Hospital and the Hansjörg Wyss Professor of Biologically Inspired Engineering at the Harvard John A. Paulson School of Engineering and Applied Sciences.

Modeling esophageal pathologies

Ingber’s and Ferri’s teams started to collaborate already in 2023 on an earlier study in which they modeled Barrett’s esophagus in a microfluidic Organ Chip. Barrett’s esophagus can be a malignant precursor of EAC, which is thought to be the result of a series of pathological changes that the epithelial lining of the lower esophagus is undergoing.

These start with inflammation, which is most commonly induced by acid reflux, and continues via the transformation of esophageal tissue into hyper-proliferating stomach and small intestine-like tissue (Barrett’s esophagus), to ultimately lead to the conversion of these highly proliferating abnormal cells into cancer cells.

Importantly, these malignant changes are not only driven by molecular and cellular processes in the esophagus’ epithelial lining, but also in its underlying “stroma,” which is made up of fibroblast cells that communicate with the cancer cells through a constant exchange of molecules, and it also contains immune cells and blood vessels.

“Whereas in our earlier work, we faithfully recapitulated the earlier stages of the pathological process potentially leading to EAC, namely Barrett’s esophagus, in our new study we fast-forwarded to its cancerous end result,” said second-author Elee Shimshoni, Ph.D., who was a Postdoctoral Fellow in Ingber’s team during both studies.

“Only by reconstituting key components of the TME and mimicking some of its fluid flows, which normally is provided by the fluid surrounding cells (interstitial fluid) and supporting blood vessels, were we able to achieve physiologically relevant drug exposure, and to accurately predict patient-specific responses to NACT in personalized EAC Chips. This could not be done using cancer organoids.”

From patients to cancer chips and back

The team engineered their TME-mimicking EAC Chip by first generating personalized EAC organoids from biopsies they endoscopically obtained from patients who were newly diagnosed with EAC but hadn’t been treated yet. First-author Sanjima Pal, Ph.D. and other members in Ferri’s team at the McGill University Health Care Centre where Ferri treats patients with esophageal cancer, had mastered the ability to create patient-matched esophageal organoids with high consistency.

Next, the team removed the organoids from the culture dish, broke them up into their constituent cells, cultured the cells in one of two parallel-running channels of a microfluidic chip the size of a memory stick, and added tumor-associated fibroblasts from the same patients to the other channel to form an adjacent tumor stroma.

Both channels are separated by a porous membrane, which allows the cancer and stromal tissues to freely exchange molecules as they would do in an actual tumor. Finally, the researchers spiked a docetaxel-based triplet chemotherapy cocktail into the nutrient fluids that flow through the stromal channel, using drug concentrations and exposure times that replicate a cycle of chemotherapy in EAC patients.

For a cohort of eight patients, all EAC Chips accurately predicted their responses to NACT within 12 days. In four of the chips, the chemotherapy caused the EAC cells to die, while in the other four chips, the EAC cells survived the chemotherapy. These results perfectly correlated with the patients’ responses to the same chemotherapy and their survival rates following surgical resection of EAC tumors.

Other authors on the study were Salvador Flores Torres, Mingyang Kong, Kulsum Tai, Veena Sangwan, Nicholas Bertos, Swneke Donovan Bailey, and Julie Bérubé.