The 1984 celebrated case of Baby Fae described the efforts of intrepid investigators at Loma Linda University in California to save the life of an infant born with a hypoplastic heart (the left side of the heart was severely underdeveloped). To salvage this unfortunate child, the heart of a baboon was transplanted into her chest, which successfully maintained her cardiac function. Unfortunately, the child died 21 days later after rejecting the heart.
The experience nonetheless spawned a lively discussion of ethics in human experimentation and the lengths to which we will go to save a life. A friend residing in New York contacted me shortly thereafter. Knowing that I was in medical training in California, he was intrigued by this report and posed the question. “What is the fastest animal in the world?” When I suggested a cheetah or African gazelle, he answered “No, a baboon running past Loma Linda Medical Center.”
I am reminded of this quip following a news report about Champions Oncology. This biotech company located in Hackensack, NJ, founded by investigators from Johns Hopkins University has developed a technique to transplant human tumors into immunocompromised (known as nude for their loss of hair) mice to test drugs and combinations. They use the term “avatar” to describe these human-tumor-bearing mice and like their video-game counterparts these mice serve as surrogates for their patient “users.” It takes 20 to 30 mice to complete the analyses for each patient. Although the work is based on sound science, the practicality, predictive validity and ultimate utility of this approach has yet to be established. This has not dampened the enthusiasm of desperate cancer patients who have proven willing to spend tens of thousands of dollars to undergo a Champion analysis.
So what are the upsides of the avatar model? For one, this is a living organism with a functioning blood supply, liver, kidneys and the capability of metabolizing pro-drugs (precursor) into active species. The process utilizes cell clusters, not individual single cells in their analysis. Responses are ultimately “phenotypic” in as much as they reflect cellular responses to injury and not genotypic profiles. Finally, toxicities can also be assessed by measuring the animal’s tolerance of the drug or combination administered.
But let’s drill down for a moment and take a closer look. The avatar approach requires months of preparation, the operation of a vivarium (animal zoo) and several mice for every single drug or combination tested. It requires prolonged (many months) maintenance of the animals with highly unpredictable engraftment of the transplanted tumors. Furthermore, significant time, energy and skill are needed to maintain these in-vivo systems.
Ultimately, only a handful of drugs can be examined, lest the number of mice required becomes unmanageable. At the end of the day these investigators are making a valiant effort to approximate work that we, and our colleagues, have successfully conducted for more than two decades – the accurate selection of chemotherapies and drug combinations for individual patients.
If we allow for the obvious downsides of expense, difficulty, time, limited sample size, low efficiency and resource intensity required to conduct even a single patient’s study there are more daunting concerns.
First, these cells are not actually in their “native state.” Over a period of time the tumors will no longer be host to human immune cells, nor will they be exposed to human cytokines and VEGF. The observed growth of the implanted tumors in the mice-avatars may, in part, reflect an ingrowth of mouse-derived fibroblasts and blood vessels, which have distinctly different biology from those of a human host.
Even if we accept the expense and difficultly of avatars, there is no clear evidence on an individual patient basis, that this approach holds any advantage over the much simpler and direct evaluation of human primary culture microspheroids. That is, the avatar approach appears to be a difficult, cumbersome, inefficient and a very expensive way to do something that we can already do inexpensively, rapidly and efficiently. Further, the purported advantages of in vivo-avatar system are actually less than meets the eye.
After all, most clinical drugs have “active” derivatives that can be utilized for testing in short-term culture without the need for a mouse liver. Our careful calibration of in vitro drug exposures against actual patient responses (P < 0.001) has established the predictive validity of these culture conditions. Finally, the toxicities of virtually all clinically relevant drugs that patients would request for testing (and likely receive) are already well-known to clinicians from existing Phase I and Phase II clinical data sets.
Human tumor sensitivity to chemotherapy (or targeted agents) is driven by what might be described as “response elements.” These unique features of each patient’s tumor can be accurately probed at the phenotypic level through the use of simple assays conducted in short-term culture. Our microspheroid model has proven highly predictive of clinical outcome in virtually every tumor type ever tested.
Our analyses are conducted in seven days, with samples that are the same size or smaller than those required for avatar generation. Furthermore, short-term platforms can analyze dozens of drugs and combinations at a price that is far less expensive. While it might be argued that avatars, once established, can be used as repositories for future research, that is small comfort to patients in need of immediate answers who find themselves paying handsomely for a service that will not be available in a timely manner, e.g. that which can help them in their need for immediate drug selection.
It seems that the medical science community is less interested in results than process. The fact that short-term cultures are predictive of clinical outcome seems less important than the provocative scientific results that these avatar models can provide Avatars enable scientists to interrogate cancer cells for genomic and proteomic signals, offering the opportunity to conduct interesting science. But has that science become more important than the clinical utility of the tests that were purportedly developed (and sold) to improve patient outcomes?
Patients who are considering spending tens of thousands of dollars for these glorified chemosensitivity tests would do themselves a service to first carefully examine the predictive validity, breadth of data, cost and turn-around-time of short-term culture methods, like the EVA-PCD® assay before they commit their precious time and resources to so “interesting” an endeavor as an avatar analysis. After all, it is the patient and their good outcome that should be at the top of the list when the advantages of any system or method are being weighed. Truly personalized cancer care should be just that – personalized.
With the rise of avatars it may be timely to re-examine the original question and wonder whether the fastest animal in the world will soon be a nude mouse running past Johns Hopkins University.