Three compounds in rosemary essential oil, α-pinene, 1,8-cineole and camphor, produced measurable but limited effects on the human serotonin transporter in laboratory tests. The study combined computer docking, a tracer-binding test and a fluorescent uptake assay in cells engineered to express hSERT. All three displaced [3H]-paroxetine and partly reduced transporter-mediated uptake, while fluoxetine produced a much stronger effect.
That pattern makes hSERT a possible molecular target for the rosemary compounds, but it does not establish that hSERT inhibition causes rosemary’s reported antidepressant-like effects. The work did not assess behavioral, neurochemical, in vivo or clinical outcomes, so it cannot establish clinical benefit.
The effects were measurable, but modest
At the strongest end of the functional experiment, the three monoterpenes inhibited uptake by 42% to 55% at 600 μM. By contrast, 0.5 μM fluoxetine inhibited about 93%. At 10 μM, inhibition was about 4% for camphor, 10% for α-pinene and 15% for 1,8-cineole. The results showed a concentration-dependent pattern, with small effects at the low end and only partial inhibition at the highest tested concentration.
A second test looked at whether the compounds displaced [3H]-paroxetine, a radiolabeled tracer used to examine binding. The concentration associated with 50% inhibition, known as the IC50, was 34.7 plus or minus 3.1 μM for α-pinene, 50.7 plus or minus 4.8 μM for 1,8-cineole and 88.1 plus or minus 7.4 μM for camphor. Fluoxetine’s IC50 was 24.9 plus or minus 2.7 nM. α-pinene was the strongest of the three monoterpenes in this test, but fluoxetine had an IC50 at a far lower concentration.
Models offered clues, not proof
Computer docking produced predicted scores rather than measured binding affinities. The approach was checked by placing citalopram back into the modeled S1 and S2 sites: the redocking RMSD values were 0.91 Å for S1 and 0.67 Å for S2, both below the 2.0 Å validation threshold. In the same model, fluoxetine scored -11.2 kcal/mol at S1 and -8.2 kcal/mol at S2. The monoterpenes had weaker predicted scores at S1: -6.0 for α-pinene, -6.0 for camphor and -5.3 for 1,8-cineole. At S2, camphor ranked highest among them at -5.6, followed by α-pinene at -5.0 and 1,8-cineole at -3.1.
The functional tests used human embryonic kidney 293 cells transiently expressing YFP-tagged hSERT. Transfection lasted 24 hours, and the cells were maintained for 24 to 48 hours before the assays. In the uptake test, cells were exposed to monoterpene concentrations from 10 to 600 μM for 10 minutes before a fluorescent probe was added for 20 minutes at 37°C. Vehicle and 0.5 μM fluoxetine served as controls. Depending on the condition, the uptake results came from three to five independent biological experiments.
What the experiments cannot answer
The central limitation is that the experiments did not identify where the monoterpenes bound. The binding and uptake results did not establish whether any compound occupied S1, S2 or another site, so the docking poses remain structural hypotheses that require direct validation. The overexpressing HEK293 system may also not reproduce the regulatory complexity of serotonergic neurons in vivo.
The high concentrations needed for the largest uptake effects add another layer of uncertainty. The study did not establish whether those in vitro effects are physiologically relevant, and it did not assess behavioral, neurochemical, in vivo or clinical outcomes. The most defensible reading is that the compounds can interact with hSERT under the study’s experimental conditions and may partially inhibit its activity in vitro. Whether that interaction contributes to rosemary-related behavioral effects remains an open question.
Paper data and sources
Original title: Interaction of major monoterpenes from Rosmarinus officinalis L. essential oil with the human serotonin transporter: molecular docking, radioligand binding, and functional uptake studies
Authors: Valle C, Segura K, Cárdenas P et al.
Journal/Repository: Frontiers in pharmacology
Status: Peer-reviewed
First online: 2026-10-10
DOI: Not available
Original paper · Full text