Optimization of Prolonged Normothermic Ex Vivo Animation of Human Tumor-bearing Liver Segments

The ability to recapitulate the complexities of solid human tumors for the purposes of drug development and testing has been, and remains, a major obstacle in the progress of cancer care. Despite great efforts expended on pre-clinical optimization using existing models, most drugs simply fail to demonstrate efficacy when subjected to phase III clinical trial scrutiny. Our interpretation of this is that currently available model systems lack the appropriate clinical predictive power, particularly because they fail to include human stromal and immune components and their intricate relationships with tumor cells. However, the potential exists for a new model of human malignancy using intact tumors removed from patients, in which the complex spatial relationships between all cell types in the tumor microenvironment are perfectly preserved. Since 2005, ex vivo organ perfusion has been used in the context of organ transplantation. However, perfusion systems have recently been used to keep whole human livers viable under normothermic conditions for one week outside of the body.

Our lab has begun optimizing our own normothermic ex vivo perfusion machine for use on tumor-bearing liver segments that are removed as part of standard treatment algorithms from patients at the NIH Clinical Center.  As a hepatobiliary oncologic surgeon, I have unparalleled access to tumor-bearing human tissue immediately after surgical resection, which minimizes ischemia prior to the organ being placed on our perfusion machine. Our machine itself recapitulates all of the major human organ systems around the resected liver. For example, we have centrifugal pumps and oxygenators that serve as our cardiovascular system, an in-line glucose monitor and insulin / glucagon infusions that serve as our pancreas, and a custom-designed diaphragm integrated with a pediatric ventilator to simulate liver movement during breathing. Most importantly, our system possesses a “central nervous system” in the form of Python code that automates many system functions. Because we have essentially “rebuilt” the human body in our system, tumor-bearing human liver segments maintained ex-vivo on our machine recapitulate the complexities of a human tumor under near-physiologic conditions. Consequently, our system serves as an ideal model to study human tumors and to use as a drug development platform. For example, real-time tumor responses to chemotherapeutic drugs delivered through the intact liver vasculature can be measured by serial biopsies taken over the course of drug delivery, which provides a window of insight into drug efficacy that would allow for the development of rational novel drug combinations.  Moreover, our system’s perfusate is amenable to continuous sampling such that biomarker discovery will be greatly facilitated. Our system allows for drug metabolism and tumor penetration to easily be measured, which is an important feature considering that this has been nearly impossible to assess in patients. Using our system, we can envision a fundamental alteration in drug development such that anti-tumor drug efficacy is evaluated first, rather than last, with subsequent efforts focused on delivery techniques and minimization of toxicity.

To date, we have maintained porcine liver viability ex vivo on our machine for up to 48 hours, and have had similar success with human livers, perfusing for up to one whole week. We are seeking the help of a summer student to help us achieve our next goal: automated maintenance of the liver during the course of study. The student will work with our multidisciplinary team consisting of surgeons, biomedical and electrical engineers, and hepatologists to further optimize the perfusion machine via:

• Development of PID algorithms in Python to regulate automated infusions of insulin / glucagon and vasoactive substances
• Integration of and in-line blood parameter monitoring system with a gas mixer to maintain physiologic oxygen saturation
• Development of an alarm system in Python to alert machine users of important changes to perfusion parameters
• Post-perfusion analysis of the liver tissue, including H&E staining, assays for free hemoglobin, bile analysis, etc.
• Help with procuring livers and monitoring the perfusion machine during runs, including taking point of care labs, serial biopsies, and blood samples

The BESIP student working in this area should have a background and interest in computer programming (experience with object-oriented programming in Python strongly preferred), biomedical engineering, human physiology, and biochemistry. Through working closely with the interdisciplinary team, the BESIP intern will gain valuable hands-on experience with multiple procedures and technologies and can expect to gain experience in various areas of cancer and biomedical research.

BETA Intern Name: Joseph Editone III

Institution: Prairie View A&M

Project Title: Impact of prolonged surgery on metastatic cancer