Beyond a cancer stem cell marker: molecular state dependent functions of CD133 and state targeted therapy

CD133 (prominin-1) is among the most widely used cancer stem cell markers, yet two decades of research have yielded persistent contradictions: CD133-negative cells can be tumorigenic, CD133-positive and -negative states can interconvert, and CD133 perturbation produces opposing phenotypes across models, while CD133 directed therapies have encountered toxicities consistent with on-target/off-tumor effects. We propose that these inconsistencies share a common origin: CD133 positivity does not define a homogeneous functional entity. Rather, the functional identity of CD133 is jointly encoded by PROM1 alternative splicing, post translational modification combinations, and subcellular localization. Depending on this molecular state, the same protein can support stemness, drive migration and invasion, sustain quiescence like states, suppress autophagy, or be exported as vesicular signals that remodel the tumor microenvironment. CD133 is therefore better understood not as a static marker with a fixed function, but as a state dependent functional integrator, in which detected expression need not reflect pathogenic function. This distinction matters therapeutically: eliminating CD133-positive cells and disrupting pathogenic CD133 states are not equivalent objectives. CD133 directed therapy should accordingly shift from marker targeting to state targeting, guided by a state resolved stratification scheme that matches isoform-, phosphorylation-, glycosylation- and localization- defined CD133 states to rational interventions.