We recently discovered a (to our knowledge) new neuroimmune interaction named the gateway reflex, in which the activation of specific neural circuits establishes immune cell gateways at specific vessel sites in organs, leading to the development of tissue-specific autoimmune diseases, including a multiple sclerosis (MS) mouse model, experimental autoimmune encephalomyelitis (EAE). We have reported that peripheral-derived myeloid cells, which are CD11b+MHC class II+ and accumulate in the fifth lumbar (L5) cord during the onset of a transfer model of EAE (tEAE), play a role in the pain-mediated relapse via the pain-gateway reflex. In this study, we investigated how these cells survive during the remission phase to cause the relapse. We show that peripheral-derived myeloid cells accumulated in the L5 cord after tEAE induction and survive more than other immune cells. These myeloid cells, which highly expressed GM-CSFRα with common β chain molecules, grew in number and expressed more Bcl-xL after GM-CSF treatment but decreased in number by blockade of the GM-CSF pathway, which suppressed pain-mediated relapse of neuroinflammation. Therefore, GM-CSF is a survival factor for these cells. Moreover, these cells were colocalized with blood endothelial cells (BECs) around the L5 cord, and BECs expressed a high level of GM-CSF. Thus, GM-CSF from BECs may have an important role in the pain-mediated tEAE relapse caused by peripheral-derived myeloid cells in the CNS. Finally, we found that blockade of the GM-CSF pathway after pain induction suppressed EAE development. Therefore, GM-CSF suppression is a possible therapeutic approach in inflammatory CNS diseases with relapse, such as MS.