(1055-D) • nELISA High-Throughput Protein Profiling of Adipocytes Provides Mechanistic Insights into Immunometabolism
Tuesday, February 6, 2024
2:00 PM – 3:00 PM EST
Location: Exhibit Halls AB
Abstract: HTS programs are increasingly adopting high-content technologies that can better inform the selection of drug candidates early on inHTS programs are increasingly adopting high-content technologies that can better inform the selection of drug candidates early on in the pipelines. While imaging-based technologies such as CellPainting provide a cost-effective and high-throughput method of capturing cell phenotypes, linking features to mechanistic insights has been challenging. In contrast, protein profiling provides ground-truth biologically relevant information, but its application to phenotypic screening has so far been limited by high costs, as well as the difficulty of scaling protein content and throughput.
We previously reported the development of the nELISA, a next-generation bead-based immunoassay platform capable of measuring 191 proteins in parallel, in high-throughput (1536 samples/day/instrument), at a cost amenable to phenotypic screening. Here, demonstrate its ability to capture metabolic phenotypes and extract mechanistic information. To achieve this, we generated adipocytes from pre-adipocytes, in the presence or absence of free fatty acids (FFA) or the inflammatory cytokines TNFalpha and IL-6. In all conditions, we screened a library of 1500 well-characterized small molecules, generating 10,368 samples. Supernatants were collected and profiled with the nELISA to quantify secreted cytokines and metabolic mediators.
Both FFA and cytokine treatment affected adipocyte secretomes, though the effects of the latter were more pronounced. Thus, cytokine treatment led to the induction of proteins unaffected by FFA, including Activin A, CCL7, and TSLP, as well as a greater induction than FFA of CXCL1, CXCL6, and IL-8. Of note, several compounds exacerbated the inflammatory effects of cytokine treatment. The expression of all of these proteins was further induced by prostanoid receptor agonists: alprostadil, bimatoprost, dinoprost, dinoprostone, iloprost, latanoprost, and travoprost. In line with these observations, hierarchical clustering of nELISA data revealed that these compounds clustered very closely, as did the seemingly unrelated compound fosinopril. Interestingly, fosinopril is an ACE inhibitor that indirectly leads to an induction of prostaglandins, and also led to increased Activin A, CCL7, CXCL1, CXCL6, IL-8 and TSLP in our dataset, further validating the exacerbating effect of prostanoid receptor activation on TNFalpha + IL-6 treatment of adipocytes. However, cyclooxygenase/prostaglandin inhibitors included in our screen did not reverse the effects of cytokine treatment, indicating that cytokine treatment was not acting through prostaglandins, but rather synergizing with them, consistent with the distinct receptors but overlapping signaling pathways of these molecules.
The example provided demonstrates the power of nELISA in phenotypic screens, enabling not only identification of compounds modulating a disease phenotype, but simultaneously providing insights into their mechanism of action.