AI annotated 101 thousand PET-CT scans: patients with brown fat had a lower prevalence of 92 diagnoses, and losartan was tested in cells and mice
AI annotated 101 thousand PET-CT scans: patients with brown fat had a lower prevalence of 92 diagnoses, and losartan was tested in cells and mice
In a Nature article published on October 14, the authors analyzed 101 243 scans from patients at Memorial Sloan Kettering, a US cancer center. An algorithm measured brown fat activity, and the team tested the association they found between losartan and this tissue in cells and mice.
Brown fat uses energy to produce heat. On PET-CT scans, radiolabeled glucose reveals areas that actively take it up, allowing researchers to assess the activity of this tissue. In a 2021 study, the same group classified 52 487 patients according to whether they had visible brown fat. The new archive allowed the researchers to measure tissue activity in different parts of the body: the software identified six such regions and calculated tissue volume and glucose uptake intensity in 8 709 patients.
In a previous mouse experiment, researchers tested a mechanism they had selected in advance: the HSF1–ADH5 molecular pathway in brown fat. Here, the clinical archive first identified candidate drugs, and experiments then examined how one of them acts on the tissue.
The authors matched patients by age, sex, body mass index, and temperature on the day of the scan, and accounted for cancer characteristics and treatment in their analyses. After these adjustments, 98 diagnoses remained statistically associated with detectable brown fat; 92 were less prevalent in patients with detectable brown fat. For type 2 diabetes, the odds ratio, which compares the odds of a diagnosis between groups, was 0,42; for chronic ischemic heart disease, it was 0,53. High brown fat activity in the armpit region was also associated with a lower prevalence of diabetes, coronary artery disease, arrhythmias, and heart failure in this cancer cohort.
The researchers used the same medical records to screen medications: of 125 drugs, 39 were statistically associated with the presence of brown fat. These included losartan and valsartan, blood pressure medications that act through the angiotensin receptor. Losartan was tested in mature brown fat cells, where it increased Ucp1 gene expression, UCP1 protein levels, and norepinephrine-dependent uncoupled respiration, a process in which cells release energy as heat.
The team then examined how the drug acts. In cells with the At1r gene disabled, Ucp1 expression still increased, even though At1r encodes the receptor that losartan usually targets. Experiments with the losartan metabolite EXP3179 and a substance that blocks PPARγ implicated PPARγ, a protein that regulates gene expression in fat cells, including expression of Ucp1.
In mice, losartan increased Ucp1 expression in brown fat and enhanced the tissue's response to a β3-agonist, a substance that mimics the signal that activates brown fat. Mice receiving the drug expended more energy when exposed to cold.