Fibroblasts are key potentiators of chronic disease pathophysiology. Despite their established roles in promoting pathological inflammation and tissue remodeling, activated myofibroblasts are generally characterized as a single, homogeneous cell population, obscuring critical functional distinctions. Defining the cell states, their molecular regulators, and restricted markers is critical to developing effective therapies for the treatment of fibrosis. Here, using a human lung stromal cell atlas of idiopathic pulmonary fibrosis, we identify two myofibroblast transcriptional states associated with distinct predicted biological function, regulation, and cell surface marker expression. We identify fibroblast-specific TGFb signaling as the key regulator of the mechanistic switch from a wound healing-associated and proliferative to a profibrotic myofibroblast. Further, we elucidate conserved TGFb-dependent and suppressed gene expression programs that define these states. Our findings reveal that LRRC15 is highly restricted to myofibroblasts that primarily express an extracellular matrix-remodeling gene program and illuminate that this key cell state can differentiate in the absence of an obligate inflammatory precursor intermediate. Last, we apply machine learning using a human single-cell foundation model to demonstrate broad applicability of the biology described herein to human chronic disease.
Keywords: TGFb; fibrosis; inflammation; myofibroblast; wound healing.