What proteomics reveals about semaglutide

A review published in the Journal of Endocrinology and Investigation on 23 July 2026 pooled 16 proteomics studies to map which proteins and biological pathways semaglutide shifts inside the body^[1]. Four of those studies looked at humans and twelve looked at animals or cultured cells, so the picture is real but still early. It is worth understanding what proteomics can and cannot tell you before reading too much into any single protein.

What proteomics actually measures

Genomics tells you which genes a cell carries. Proteomics measures the proteins those genes actually produce at a given moment, which is closer to what the cell is doing. Researchers take a tissue or fluid sample, break it down, and use mass spectrometry or antibody-based panels to count thousands of proteins at once. When a drug changes the level of a protein, that shows up as a shift in the proteomic profile. The appeal for a drug like semaglutide is that it hints at mechanism: not just that people lose weight, but which molecular machinery moves when they do.

What the review found

Across the 16 studies, the authors reported proteomic changes in the metabolic, cardiovascular, hepatic, pulmonary, renal and nervous systems, drawn from samples of serum, plasma, adipose tissue, heart muscle, aorta, lung, kidney, hippocampus and skeletal muscle^[1]. A handful of themes recurred across otherwise different experiments. Fatty acid metabolism and the PPAR signalling pathway came up repeatedly, which fits a drug that reshapes how the body handles fat. So did oxidative phosphorylation and mitochondrial function, the cell's energy production line. Inflammatory and complement protein networks shifted, as did proteins tied to extracellular matrix remodelling, the scaffolding that holds tissue together. The review also flagged mTORC1 signalling, a hub that governs cell growth and metabolism.

The studies used a mix of platforms, including SomaScan antibody panels, tandem mass tag mass spectrometry, data-independent acquisition proteomics, and specialised phosphorylation and mitochondrial workflows. That variety is a strength when different methods point to the same pathway, because it lowers the chance the signal is an artefact of one technique. It is a weakness when they do not agree, because there is then no clean way to reconcile the numbers.

What this review does not tell you

The authors are direct about the limits, and readers should be too. Most of the evidence is animal or in-vitro work, with only four human population studies in the pool. Sample sizes were small, the people and animals studied differed a lot from one another, and the proteomic platforms varied enough that results are hard to line up side by side^[1]. A protein moving in a mouse heart does not automatically mean the same thing happens in a person, and a protein that shifts in the blood is a correlation until a trial shows it changes an outcome that matters. None of this is a knock on the work. It is the normal state of an early mechanistic field, and treating a pathway signal as a proven clinical benefit is exactly the mistake the review warns against.

How this fits with what we already know

Semaglutide is the most heavily trialled peptide in our library, so the proteomics sits on top of a solid outcomes record rather than replacing it. In the STEP-1 obesity trial, participants on semaglutide 2.4 mg lost a mean of 14.9% of body weight at 68 weeks, against 2.4% on placebo^[2]. In the SELECT trial, adults with overweight or obesity and established cardiovascular disease but no diabetes had a 20% lower rate of major adverse cardiovascular events, a hazard ratio of 0.80^[3]. Those are hard endpoints measured in tens of thousands of people. The proteomics work tries to explain the machinery underneath them, which is useful, but it is a different and earlier kind of evidence.

For now the honest read is that proteomics is starting to sketch how semaglutide acts across more than one organ system, and the recurring hits on fat metabolism, mitochondria and inflammation are consistent with its known effects. Whether any specific protein becomes a useful marker for who responds, or who is at risk of a side effect, is an open question that larger human studies will have to answer. Semaglutide remains a prescription-only medicine everywhere we cover, and none of this research changes that. See our semaglutide page for the regulatory status where you live, and treat this as an educational summary rather than medical advice.