When making diagnoses, doctors have long followed their noses. Hippocrates noted the body odor of his patients when identifying ailments. More than a millennium later, the Persian physician Avicenna used the smell of a person’s urine to detect illness. And in the early age of the house call, physicians made sniffing patients part of their routine.
At first whiff, such practices might reek of superstition, but intuitively and scientifically, they make sense. The bodies of sick people produce volatile chemicals that are identifiable by smell. People with diabetic ketoacidosis, for example, often have fruity-scented breath, while those with chronic kidney failure emit a fishy odor.
Hoping to make use of these chemical fingerprints, researchers are exploring ways to sniff out disease. Some are studying the abilities of cancer-sniffing dogs, which have been shown to diagnose illness with surprising success. Others have spurned the hairy and unpredictable business of doggy diagnostics in favor of artificial noses, hoping to develop noninvasive devices for detecting disease. Although some devices are now being tested, their development is stunted by a fundamental lack: Relatively little is known about our sense of smell.
Enter Rachel Wilson, an HMS associate professor of neurobiology and one of a growing number of scientists now focusing on olfaction. Often considered the neglected stepsister of sensory research, olfaction has grown substantially as a field since the first olfactory receptor genes were discovered in 1991. That discovery opened the door for Wilson and others to study our sense of smell at the molecular level.
Wilson’s team studies the relatively simple olfactory system of the fruit fly, Drosophila melanogaster. The team investigates how the nervous system collects olfactory information, and how that information is processed as it moves from the primary sense organs to the decision-making centers of the brain.
In insects, the primary sensory organs are olfactory receptors on the antennae (as opposed to receptors in the mammalian nose). Wilson’s team puffs odors at the antennae and, with the help of fine electrodes, makes electrical recordings of the neuron activity both in the receptors and in the brain as information is passed along for processing. By labeling specific groups of neurons with green fluorescent protein, and by manipulating or knocking out certain populations of neurons, the team is able to see how those neuron groups respond when fruit flies encounter odors. As the scientists tinker with each distinct group, they can begin to make inferences about the roles those neurons play in the olfactory system.
“We’re asking a number of basic questions,” Wilson says. “How does a neuron respond to odors? What odors does it respond to? How does it help to encode the olfactory world? And why does it respond that way?”
Answering such questions will help researchers understand how the nervous system identifies scents, but it won’t necessarily explain our more complicated responses to odors. With that in mind, Wilson’s lab has branched into behavioral experiments. In these studies, they observe how fruit flies respond to odor stimuli. They monitor flies in flight, for instance, to see how flight patterns change as the flies detect smells in the air. The researchers also spy on the flies as they navigate to competing odor sources.
Ultimately, complex behavioral questions such as these may prove most relevant to the development of artificial noses. Learning how the brain identifies smells should be useful not only for designing devices that can diagnose illness, but also for recognizing and tracing odors to their sources, important goals in fields such as environmental quality management and law enforcement.
Wilson believes that the differences between insect and mammalian anatomy can benefit artificial nose research. Her lab has found, for example, that fruit flies respond rapidly to airborne odor filaments. These thread-like plumes emanate from odor sources, much like smoke plumes rising from a candle. Since insects smell with their antennae, they gather information from these odor filaments on contact. As they fly into the plumes, the fruit flies alter their flight path and wing-beating rate, actions that tell researchers that the insects have detected both an odor and its source.
Mammals, in contrast, detect odors through respiration. Not only do we dilute and mix odors when we inhale, but we also disrupt the integrity of odor filaments. (Imagine taking a suction cup to that plume of smoke!) Given that the concentrations of these filaments provide clues about the direction of an odor source and its distance, the mammalian system seems comparatively ill equipped for odor detection and navigation. That’s why Wilson is surprised that more developers aren’t thinking in terms of the insect system when conceiving of artificial noses.
“Artificial nose developers are designing systems that tend to draw air in through the device’s sensors, as if it’s respiration,” Wilson says. “They should be working in parallel on systems in which sensors respond to filaments quickly, without creating turbulence, because you can learn a lot by observing those filaments.”
Wilson predicts medicine’s return to the olfactory sense as a diagnostic tool. “It has just seemed too uncouth to be sniffing patients,” she says, referring to past practices. “But to avoid using our sense of smell is to throw away valuable information—and real potential applications.”