For a century, scientists have turned to mice in order to understand our biology—not just our common biology as mammals, but also our specifically human biology. Researchers have made major efforts to create lab mice that replicate some essential aspect of a human disease. Now a recent technological breakthrough makes it possible to “humanize” mice to an unprecedented degree, but it also raises the question: How many of our human traits can we build into a mouse?
Why do researchers use mice? There are the obvious reasons: they are small and easily kept; they breed quickly; and as mammals, they share much of our biology. While mice obviously aren't tiny humans, for many purposes they are genetically close enough. The broad outlines of the mouse and human genomes are very similar, and nearly all of our roughly 21,000 genes have a murine counterpart. Because mice are so similar to us, researchers can often take a disease mutation discovered in humans and make the corresponding mutation in a mouse in order to learn something about the molecular underpinnings of the disease. There are now many mouse genetic "models" that exhibit key features of human diseases, including common diseases like heart disease, obesity, asthma, type I and type II diabetes; cancers, including Leukemia, pancreatic cancer, and melanoma; and neurological diseases such as autism, Huntington's disease, Parkinson's disease, and schizophrenia.
How many of our genes can function properly in a mouse? How many mouse genes can you simultaneously replace with human ones and still get a viable mouse? We have no idea.
These mouse models have proven tremendously useful in basic research, but new disease treatments that are developed in mouse studies have a shockingly high failure rate—80 percent—when they are tried in humans. A fair number of these failures happen because the mouse studies were poorly done, but often the problem is that the biology of these mice isn't human enough. To overcome this issue, scientists attempt to "humanize" lab mice: Instead of simply mutating the mouse version of the gene being studied, researchers conduct a direct DNA transplant to completely replace the mouse gene with the human version. You might think that, like an attempt to run an Android app on an iPhone, humans genes wouldn’t work so well in the mouse operating system. But humanizing mice with DNA transplants works surprisingly well. Last week, three different teams of scientists reported the largest successful human-mouse DNA transplants to date.
The first research team, based at Yale University and University Hospital Zurich, set out to create a mouse that could properly transform human blood stem cells into fully functional immune cells, a process that is a critical part of a healthy immune response. To study this aspect of human immunity, scientists frequently inject human blood stem cells into mice and observe them as they develop into mature immune cells. Unfortunately, the mouse immune system components don’t always play nice with the human cells. By swapping four mouse genes with their human versions, the researchers created mice with humanized immune systems that properly mature human blood stem cells.
The other two research teams, one at U.K.'s Sanger Institute in the U.K. and the biotech company Kymab, and the other at Regeneron Pharmaceuticals in New York, swapped out a different set of immune system genes to create mice that make human antibodies. Both groups, using somewhat different technologies, substituted large human antibody genes for the mouse versions. The human genes appear to function just fine in the mice, which can now produce therapeutic antibodies suitable for treating human disease. More significant, though, are the technological implications: One of the teams replaced 0.002 percent of the mouse genome with human DNA. That may not sound like much, but it's the largest replacement ever carried out in a mouse, and it's a game-changer.
THE ABILITY TO MANIPULATE and transfer very short bits of DNA revolutionized molecular biology in the 1970s by removing the barriers to previously impossible experiments. The latest technologies are tearing down a new set of barriers. Earlier this month, a team of researchers reported the first complete synthesis of a "designer" chromosome. In this case it was a yeast chromosome, and thus small by human standards, but it was a thousand times larger than what researchers get from the current standard technology for DNA synthesis. It's not cheap, but we now have the technology to replace each mouse gene with its human counterpart, and thereby ask previously impossible questions. How many of our genes can function properly in a mouse? How many mouse genes can you simultaneously replace with human ones and still get a viable mouse? We have no idea.
A way to rephrase these questions is this: How modular are human traits? What chunks of our biology can we isolate and transfer to mice, in order to better understand ourselves?
An example of the possibilities and limits of modularizing our biology is a 2009 study (PDF), where researchers at the Max Plank Institute in Leipzig humanized a mouse gene that plays an important role in an essential human trait: speech. Mutations in the gene FOX2P in humans result in severe impairments in the ability to speak and process grammar. Two small differences in FOXP2 distinguish the human version of this gene from that of other mammals, including chimps, our closest non-speaking relatives. It's tempting to see these two small changes in a single gene as somehow central to one of our most human traits. The Leipzig researchers made those two changes to the mouse FOXP2, in the hope that the results would lead insight into the evolution of human speech.
But the results were more tantalizing than illuminating. The humanized FOXP2 mice showed distinct changes in the sounds they made and in the development of certain neurons. But it's difficult to say much about the evolution of a complicated trait like language from the observation that "medium spiny neurons have increased dendrite lengths and increased synaptic plasticity" in humanized FOXP2 mice. Divorced from their native context, single human genes are unlikely to reproduce much of our biology in mice, although such experiments can generate important clues. But now we can swap in multiple human genes at once and recreate whole biological systems in mice. How human can we make lab mice? We don't know what the limits are, but, given the new technology coming online, we haven't reached them just yet.