Abstract The 18 kDa translocator protein (TSPO) is a classical marker of neuroinflammation targeted for in vivo molecular imaging. Microglial cells were originally thought to be the only source of TSPO overexpression but astrocytes, neurons and endothelial cells can also up-regulate TSPO depending on the pathological context. This study aims to determine the cellular origin of TSPO overexpression in a simplified model of neuroinflammation and to identify the molecular pathways involved. This is essential to better interpret TSPO molecular imaging in preclinical and clinical settings. We used lentiviral vectors (LV) to overexpress the ciliary neurotrophic factor (CNTF) in the right striatum of 2-month-old Sprague Dawley rats. A LV encoding for β-Galactosidase (LV-LacZ) was used as control. One month later, TSPO expression was measured by single-photon emission computed tomography (SPECT) imaging using [ 125 I]CLINDE. The fluorescence-activated cell sorting to radioligand-treated tissue (FACS-RTT) method was used to quantify TSPO levels in acutely sorted astrocytes, microglia, neurons and endothelial cells. A second cohort was injected with LV-CNTF and a LV encoding suppressor of cytokine signaling 3 (SOCS3), to inhibit the JAK-STAT3 pathway specifically in astrocytes. GFAP and TSPO expressions were quantified by immunofluorescence. We measured a significant increase in TSPO signal in response to CNTF by SPECT imaging. Using FACS-RTT, we observed TSPO overexpression in reactive astrocytes (+ 153 ± 62%) but also in microglia (+ 2088 ± 500%) and neurons (+ 369 ± 117%), accompanied by an increase in TSPO binding sites per cell in those three cell populations. Endothelial cells did not contribute to TSPO signal increase. Importantly, LV-SOCS3 reduced CNTF-induced astrocyte reactivity and decreased global TSPO immunoreactivity (-71% ± 30%), suggesting that TSPO overexpression is primarily mediated by reactive astrocytes. Overall, this study reveals that CNTF induces TSPO in multiple cell types in the rat striatum, through the JAK2-STAT3 pathway in astrocytes, identifying this cell type as the primary mediator of CNTF effects neuroinflammatory processes. Our results highlight the difficulty to interpret TSPO imaging in term of cellular origin without addition cellular analysis by FACS-RTT or quantitative immunostainings. Consequently, TSPO should only be used as a global marker of neuroinflammation.