When Joshua Woolley’s team at the University of California, San Francisco registered a small psilocybin trial for people with Parkinson’s disease, the blood draws went into the fine print. The registry entry lists them among the study’s “other” outcomes and says the samples were collected “for archiving and eventual analysis,” with that analysis “to be supported by another funding source.” Between August 2021 and the end of 2024, blood came from 9 of the 12 participants, drawn the day after their first dose and again 1 and 30 days after their second. Then it sat in the bank, waiting on someone else’s money.
| Measure | Part | Whole |
|---|---|---|
| participants gave blood samples | 9 | 12 |
On Wednesday the wait ended. A research letter in JAMA Psychiatry, led by UCSF psychiatrist Ellen Bradley and Columbia University’s Parker Kelley, reports that five inflammatory signaling molecules in those samples fell after two psilocybin sessions and were still significantly below baseline a month after the final dose. A blood marker of mitochondrial stress improved sharply right after dosing, then lost the gain by the longer follow-up. The authors write that, to their knowledge, “no published psychedelic study has measured mitochondrial markers in humans,” and that earlier human studies of psychedelics and the immune system gave mixed results, none of them in people with Parkinson’s or any other neurodegenerative disease.
The targets were chosen for a reason. Parkinson’s is known for its dying dopamine neurons, but it is also an inflamed disease. A 2022 review in Nature Reviews Immunology by Malú Tansey and colleagues notes that pro-inflammatory cytokines “are elevated in the serum of patients with PD and correlate with disease severity and disability.” It also describes mouse work in which animals lacking the Parkinson’s genes Pink1 or Prkn lost dopamine neurons under mitochondrial stress, in a process driven by DNA leaking out of damaged mitochondria. Inflammation and mitochondrial trouble are two of the processes implicated in the disease beyond dopamine itself, and the letter went after markers of both.
The samples came out of a pilot built to test safety, not immunology. As published last year in Neuropsychopharmacology, seven men and five women with mild to moderate Parkinson’s and depression or anxiety took 10 milligrams of psilocybin, then 25 milligrams about two weeks later, with psychotherapy around each session. On each dosing day, ten of the twelve had treatment-emergent adverse events. Two-thirds were anxious on the 10-milligram day, half had nausea and half a rise in blood pressure on the 25-milligram day, and two had severe anxiety during a session. One reported more thoughts about medically assisted dying a month later, which the authors did not count as a serious adverse event. Depression scores on the MADRS, which averaged 21 at baseline, were 9.3 points lower three months out, and patient-reported motor difficulties fell 7.5 points at one month, while the clinician-rated motor exam moved 4.6 points with the weakest effect size of the three. “We are still in very early stages of this work, but this first study went well beyond what we expected,” Bradley said in a UCSF release at the time, where the team floated calmer inflammation as one possible explanation. This week’s letter is the first attempt to measure it in these patients.
Nine people, no comparison arm
The design sets the limits. There was no control group, and every participant knew they were getting the drug. Inflammatory markers shift with sleep, stress, infection and mood, and these were people whose depression was lifting while they received hours of therapeutic attention, so a single-arm study has no way to separate psilocybin from everything delivered alongside it. The registry set out to examine interleukins, tumor necrosis factors and C-reactive protein, “including but not limited to.” Nine sets of samples cannot say which of those moved because of the molecule.
The investigators don’t claim otherwise. “This gives us an initial biological signal to investigate,” Bradley told NPR. Aoife O’Donovan, a UCSF psychiatry professor and co-author, said the blood findings need confirmation in follow-up studies before anyone makes definitive claims, while pointing out that they are “objective markers,” unlike the symptom scales that carried the parent trial.
The two markers kept different clocks. The inflammation stayed down. The mitochondrial marker did not. That divergence proves nothing by itself, since different biomarkers can rise and fall at different speeds under the same influence, mood included. What it offers a properly controlled trial is something specific to test: a brief effect on cellular stress and a slower, more durable one on the immune system, as against a single undifferentiated glow of feeling better.
Who gains if it holds
Psychedelic medicine is a business now, and its research gets read the way any industry’s does. The parent pilot was paid for by an anonymous donor and used psilocybin developed by the Usona Institute, and the 2025 paper’s disclosures put several investigators close to companies in the field. O’Donovan has received research support from SILO Inc. and Filament Health. Woolley has received research support from Usona and Filament and has consulted for MindMed, among others. Andrew Penn, a co-author, has consulted for Compass Pathways and MindMed and holds shares in Tactogen and Osmind. Bradley declared no competing interests. A research grant is not a consulting fee, and neither is equity, and none of it invalidates nine people’s blood work. It does mean that some of the people describing the signal have commercial relationships in an industry whose market grows with every new indication.
The incentives also cut the other way. Charles Nichols, a professor at LSU Health Sciences Center in New Orleans who was not involved in the study, called the letter “really exciting” and said it “validates a lot of the pharmacology” his group has worked on. About 15 years ago his lab found that a psychedelic acting on the 5-HT2A serotonin receptor was a “potent anti-inflammatory” in cell culture modeling blood-vessel inflammation, an effect NPR reported came at doses well below those that produce a psychoactive experience. His group has since run experiments in animal models of asthma, arthritis and cardiovascular disease. Nobody has shown anything of the kind for low-dose psilocybin in people. If the immune effect ever turns out not to need the trip, though, its value would sit apart from the hours-long supervised session on which much of the psychedelic-therapy business has been built.
In 2025 UCSF announced a randomized controlled trial to follow the pilot, with a second site at Yale, a goal of 100 participants, and funding from the same anonymous donor and the Michael J. Fox Foundation for Parkinson’s Research. The comparison arm missing from the first study is built into that one.
The registry entry that first promised the blood analysis still lists no collaborators. It never says who the other funding source turned out to be.
Sources
- JAMA Psychiatry: Bradley, Kelley et al., “Psilocybin and Inflammatory and Mitochondrial Stress Markers in Parkinson Disease: A Nonrandomized Clinical Trial” (2026)
- ClinicalTrials.gov registry entry NCT04932434
- Neuropsychopharmacology: Bradley et al., open-label psilocybin pilot in Parkinson’s disease (2025)
- NPR via OPB: New human data hints at psychedelics’ potential to combat inflammation (2026)
- Nature Reviews Immunology: Tansey et al., “Inflammation and immune dysfunction in Parkinson disease” (2022)
- UCSF: “How ‘magic mushrooms’ could help Parkinson’s disease patients” (2025)