Out of Order: Modestly modeled
Randall C Willis talks about what's wrong with the disease models we currently use to discover and develop therapeutics and points us in the general direction of where we need to go to get models more in line with reality
Register for free to listen to this article
Listen with Speechify
0:00
5:00
For a brief period of my childhood, I dabbled in model airplanes and model ships. And by “dabbled,” I mean I spent an inordinate amount of time with my fingers glued together. But aside from the medical agonies of modern chemistry, what struck me most about the exercise was how pale an approximation these models were of the real thing—about the only thing my miniature Spitfire had in common with the WWII fighter was the sheer carnage of the plane as it went from airborne to groundborne in its flight across the room.
More recently, my fascinations have turned to models of a more human variety, such as those found on the catwalks of America’s Top Model. (Let’s face it, after DDNews, I am all about the latest issue of Vogue.) And like the plastic variety described above, these models seem to be at best a glittery approximation of the actual thing.
It’s often hard to believe that I share about 100 percent DNA sequence identity with these mystical creatures. Now, 98 percent coherence with bonobo chimps, I have no problem believing.
All this to say that we are constantly surrounded with models that are poor facsimiles of the real deal upon closer inspection. So, it should probably should come as no surprise that the same is true in medicine and drug discovery.
Last week, I read a story in the newspaper that touted the life-prolonging properties of the diabetes drug metformin; a regular fountain of youth, the headlines implied. And yet, after coursing through the article, I eventually discovered that the rejuvenation experiments were performed on the worm C. elegans, which was shocking for two reasons.
One, I did not realize that there were largely untapped market opportunities in nematode diabetes. And two, the life-extending implications were being made based on a species that wasn’t even a chordate, let alone a mammal.
Now, I appreciate that this is an extreme example, probably overhyped by an eager press officer, but the literature is rife with examples of models that completely failed to live up to expectations when researchers tried to match success in model systems with success in actual humans.
As oncology god Judah Folkman once mused, we have become really good at curing cancer in mice.
Part of the challenge, I think, is that because we cannot experiment directly on humans, or at least not within the editorial reach of DDNews, researchers are often forced to study new compounds or therapeutic modalities on approximations of approximations of approximations of human disease.
We don’t study the impact of irinotecan on human colorectal cancer but rather extrapolate from its effects on an induced form (approximation 1) of murine colon cancer (approximation 2) in mice (approximation 3). Or we study new biologics against a chemically induced inflammation in dogs that bears a passing resemblance to rheumatoid arthritis in humans.
Within the realm of in-vitro models, the advent of technologies like 3D cell culture and microtissues is adding some biological context back into the completely artificial realm of 2D cultures of immortalized cell lines. (For more on this, see the special report “Life moves on” in this issue, on page 21). But in the absence of factors such as tissue vascularization and the like, even these advances result in weak approximations.
The goal of a better, more representative model may be getting a step or two closer, however, with the help of stem cells.
As we’ve reported previously in these pages, and as I am presently hearing at the ISSCR conference in Vancouver, stem cells are giving us enhanced opportunities to study human disease generated from the source material—human cells—potentially down to the scale of the individual patient.
The standard technical limitations of in-vitro analysis hold for stem cells—a lump of microtissue in a microwell dish does not a micro-human make—but we do have the opportunity to limit one or two approximations.
At ISSCR, for example, Daniela Cornacchia and colleagues at Sloan-Kettering and Weill Cornell Medical College describe their efforts to understand the inadvertent de-aging of cells transformed into iPSCs. Even when taken from older patients, the reversion process makes it difficult to use the cells to study late-onset diseases. The group is trying to identify factors that will allow them to induce natural aging into these cultures to improve models of such diseases as dementia.
Similarly, Rohan Nadkarni and Carlos Pilquil of McMaster University are endeavoring to produce 3D lung tissue from iPSCs that contain both conducting and gas-exchange zones mimicking normal lung function. If the model bears out, it may provide an even more realistic platform to study respiratory diseases in vitro.
Until we are in a position where we can do high-throughput human screening—in a 96-well cube farm, perhaps—the search for better model systems must be a priority. And given the challenges of translating preclinical success into clinical success, perhaps it should be a higher priority than the development of new therapeutics.
Look for more on this topic in a special feature on disease modeling in the November issue of DDNews.
