American Scientist

In 2009, chemist Perdita Barran of the University of Manchester received what turned out to be a momentous phone call from her collaborator, Tilo Kunath of the University of Edinburgh. A 68-year-old grandmother and nurse, Joy Milne, had stood up at a public talk Kunath was giving about his research on stem cells and Parkinson’s disease and asked why people with the neurodegenerative disorder had a distinctive smell. That possibility wasn’t one Kunath had considered before, but Milne was very convincing. “Her question had obviously affected him,” Barran says, remembering the phone conversation. What Kunath wanted to know: Could Barran identify what Milne was smelling? “I said, ‘That’s what I do,’” Barran recalls. “I explained that smells are molecules, and we have receptors in our brains that ‘see’ these molecules and turn them into signals. As a mass spectrometrist, I measure molecules by their weight.”

After 10 years of experimentation, Milne, Kunath, and Barran have their answer—a team led by Barran published their results in March in ACS Central Science—and it may lead to a noninvasive screening test for early detection of Parkinson’s disease.

About 1 in 15 people will be affected by Parkinson’s disease in their lifetimes. Symptoms of Parkinson’s disease include tremors, rigid muscles, speech changes, and sleep disruption. The number of people with Parkinson’s disease doubled between 1990 and 2015 to more than 6 million. And that number is predicted to double again by 2040. Parkinson’s patients may struggle for months or years without a diagnosis and proper treatment, especially the 1 in 40 people who develop the disease before the age of 45. Because the disease is associated with the elderly, younger Parkinson’s patients often struggle a long time before they are diagnosed. That was true for Milne’s late husband, Les.

For Milne, the story goes back to 1982, when Les, then in his early 30s, began to smell different. She noted it, but didn’t think much of it. In 1994, when her husband was 44, he was diagnosed with Parkinson’s disease as his symptoms became more obvious. But it took another 15 years before Milne associated the smell with Parkinson’s. In 2009, she and her husband got involved with a group for people with the neurological disorder, and she noted that all the people she met with the disease smelled like her husband. Because he was an anesthesiologist and she is a nurse, they knew how to find a researcher to approach about their hypothesis.

When Barran received the call from Kunath, the first thing she thought was that Milne was smelling age-related changes. “It’s my eternal shame, but I’ve redeemed myself now,” Barran says. Still, her skepticism was a guide: They knew they needed to overcome these obvious questions about whether Milne was misattributing what she smelled.

Barran’s team devised a simple experiment first: They found six people with Parkinson’s and six healthy people as controls who were about the same age. The study participants wore T-shirts and then gave them to Barran’s team, who cut them in half and then did a blind test with Milne to see if she could correctly identify the people with Parkinson’s by smell. She was able to correctly and consistently attribute all the T-shirts to people with Parkinson’s.

But Milne also said one of the T-shirts from the control group smelled like the ones worn by people with Parkinson’s. This false positive initially was a problem. The group wasn’t able to convince funders that the experiment was worth doing—that is, until that control group member contacted them eight months later to let them know he had been diagnosed with Parkinson’s disease.

Barran still needed to home in on the source of the smell, and the T-shirt experiment had offered an important clue. “We had thought that the smell would be strongest in the sweat in the armpit,” Barran explains. “But it was strongest in the part of the T-shirt that sits under the neckline, where there is a lot of sebum, which is the oily substance that we excrete all over our bodies except on the palms of the hands and the soles of the feet.” That clue guided their clinical trial, which sought to detect the molecules Milne was smelling.

For each of the study participants, 43 with Parkinson’s and 21 control participants, the team swabbed with gauze areas of the body where sebum builds up. They then heated up those samples and put them through a mass spectrometer, which separates the molecular components by mass. Barran’s team set up an “odor port” through which Milne could smell the separated molecules as they passed through the equipment and note when she smelled the scent she associated with Parkinson’s patients.

This study isolated 17 molecules that were different when comparing sebum from patients with Parkinson’s and sebum from the control group. Nine of those were validated using an independent cohort of 31 participants; four of those nine molecules passed through the odor port at a moment when Milne confirmed she was smelling an odor she associated with Parkinson’s. Those four compounds alone can predict Parkinson’s disease in 9 out of 10 people, making these molecules potential biomarkers for Parkinson’s disease.

“The identification of biomarkers is an important topic in disorders that are often difficult to diagnose and treat,” says Vincenzo Angelo Donadio, a neurologist at the IRCCS Instituto delle Scienze Neurologiche in Italy, who was not involved with the study. “Reliable biomarkers may help the clinician improve the clinical management of patients affected by these disorders.” However, he cautions that the small sample size may limit the salience of Barran’s findings. Also, signs and symptoms of Parkinson’s disease can be caused by multiple neurodegenerative disorders associated with various potentially interacting hallmark proteins. “The lack of analysis of different parkinsonisms may prevent establishing how specific the result of this test is in identifying patients with Parkinson’s disease,” Donadio says. “This study must be considered preliminary.”

Barran is continuing to research how useful these molecules are for screening for Parkinson’s disease. And because the biomarkers smell stronger when they are in sebum, she thinks she could find biomarkers for other diseases in this substance on the skin. “The potential to use sebum as a kind of sampling medium for diagnostic systems is really exciting,” Barran says.

The results also raise questions about why these molecules are present in greater quantities in the sebum of people with Parkinson’s. The molecules identified were metabolites. Whether they came from the human body or from microbes on the skin is unknown; ongoing work on the skin and gut microbiome of Parkinson’s patients could lend clues. Regardless, Barran, Kunath, and Milne are staying hot on the trail of this intriguing scent.