In a recent study published by Bai et al in Science Translational Medicine, researchers detailed the mechanisms by which some patients with rheumatoid arthritis experience pain without having synovial inflammation. Prior research has shown that pain scores reported by patients with rheumatoid arthritis may not correlate with synovial inflammation. The researchers developed a machine-learning model—known as the graph-based gene expression module identification—to examine synovial biopsy samples from patients with rheumatoid arthritis who had limited synovial inflammation. They found that 815 genes were linked to pain scores in those with established rheumatoid arthritis, and that most of these genes were present in patients with early or untreated rheumatoid arthritis who had low synovial inflammation. After performing single-cell and single-nucleus RNA sequencing, the researchers discovered that these genes were expressed predominantly in the synovial lining fibroblasts—which were thought to interact with dorsal root ganglion nociceptors expressing calcitonin gene–related peptide. In culture treated with human synovial fibroblast supernatant or netrin-4 protein, mouse dorsal root ganglion neurons demonstrated increased neuron branching. Additionally, in human synovial hypertrophic papilla, the researchers observed little inflammation and calcitonin gene–related peptide–expressing nociceptive axons encasing blood vessels. The researchers hope to use these findings to develop novel therapies to reduce pain in patients with rheumatoid arthritis and low levels of synovial inflammation.