Doctors and soldiers could soon place their trust in an unusual ally: the mouse. Scientists have genetically engineered mice to be ultrasensitive to specific smells, paving the way for animals that are “tuned” to sniff out land mines or chemical signatures of diseases like Parkinson’s and Alzheimer’s.
Trained rats and dogs have long been used to detect the telltale smell of TNT in land mines, and research suggests that dogs can smell the trace chemical signals of low blood sugar or certain types of cancer. Mice also have powerful sniffers: They sport about 1200 genes dedicated to odorant receptors, cellular sensors that react to a scent’s chemical signature. That’s a few hundred less than rats and about the same as dogs. (Humans have a paltry 350.)
Paul Feinstein wants to upgrade the mouse’s already sensitive nose. For the last decade, the neurobiologist at Hunter College in New York City has been studying how odorant receptors form on the surface of neurons within the olfactory system. During development, each olfactory neuron specializes to express a single odorant receptor, which binds to chemicals in the air to detect a specific odor. In other words, each olfactory neuron has a singular receptor that senses a particular smell. Normally, there is an even distribution of receptors throughout the system, so each receptor can be found in about 0.1% of mouse neurons.
Feinstein wondered if he could make the mouse’s nose pay more attention to particular scents by making certain odorant receptors more numerous. He and colleagues developed a string of DNA that, when injected into the nucleus of a fertilized mouse egg, appears to make olfactory neurons more likely to develop one particular odorant receptor than the others. This receptor, called M71, detects acetophenone, a chemical that smells like jasmine. When the team added four or more copies of the DNA sequence to a mouse egg, a full 1% of neurons carried it—10 times more than normal.
In the same lab, Hunter College postdoctoral researcher Charlotte D’Hulst was trying in vain to splice human olfactory genes into a mouse’s genome. For some reason, a tried-and-true gene swapping method just wasn’t working. So she teamed up with Feinstein and used his technique to introduce a human odorant receptor, OR1A1—modified with Feinstein’s DNA sequence—into mice. OR1A1 detects a chemical with a peppermintlike smell.
It worked. And this time, they found the OR1A1 receptor in a whopping 13% of the olfactory neurons. “We changed the probability of choice in our favor,” D’Hulst says, “even though we still don’t really know how this DNA sequence drives choice.”
To figure out how these changes affect a mouse’s sense of smell, the researchers gave both the genetically engineered and ordinary mice a choice of two water bottles, one of which was diluted with trace amounts of chemicals that triggered either M71 or OR1A1. If the mice drank from the pure water, nothing happened, but if they drank from the diluted water, they would receive an uncomfortable injection that made them feel queasy. The researchers kept lowering the concentration of the chemicals to see if the mice could detect, and therefore avoid, it. Both types of super-sniffer mice were able to detect the odors at much lower concentrations, preferentially seeking out the pure water. The M71 mice were about twice as sensitive and the OR1A1 mice about 100 times as sensitive to their respective chemicals as were ordinary mice, the researchers write today in Cell Reports.
Feinstein predicts that with more tweaking, his introduced DNA sequence can achieve even greater sensitivity—to a limit. “We’re not sure where the break point is, the point of diminishing returns,” he says. “But I believe we can increase it further.”
The work looks “very encouraging”, says Alexander Fleischmann, a neuroscientist at the Collège de France in Paris who also studies odorant receptors in mice. But he wants to know whether the technique holds up across a broader range of settings. For example, would mice exhibit different behaviors if they were tempted with rewards for scent detection, rather than punished with an uncomfortable injection as they were in this study? It’s more than an academic question. Behavioral conditioning that relies upon punishment, he says, involves split-second reactions in the brain that differ from the situations mice would face in a real-world scenario.
There’s also a question of signal-to-noise. More receptors also means more olfactory signal going to the brain, and in his own work, Fleischmann has seen that the brain’s olfactory bulb tends to flatten out spikes in odor signals. That might potentially limit the effectiveness of boosting the proportion of a particular receptor.
Meanwhile, Feinstein says the new technique could help answer bigger questions, including decoding the “black box” of the human olfactory system—so called because so little is known about how the human brain processes smells. If the researchers could make a mouse for each human odorant receptor—their stated goal—they could discover which chemicals trigger each receptor, something only imperfectly understood today. “We believe that by using this technique, we finally have the tools to crack the olfactory code,” he says.
There are practical applications as well. The work could help engineers create a “nose-on-a-chip,” for example, that would allow, say, fragrance manufacturers to more precisely tailor their scents. Feinstein also envisions translating the technique to rats to create supersniffing land mine detectors able to detect TNT at extremely low concentrations, which would allow them to find well-hidden or disguised dangers. In addition, the biosensors could also be used to detect trace chemical signatures of diseases.
“It doesn’t have to be a smell,” Feinstein explains. “You can get biosignatures for diseases like Parkinson’s, Alzheimer’s, or tuberculosis—anything that causes a chemical change in our bodily fluid that can be detected.”