Melatonin exerts colony-stimulating activity and rescues myeloid progenitors from apoptosis, induced either in vivo or in vitro by cancer chemotherapy compounds in tumor-bearing mice. These effects are mediated mainly by T-helper cell-derived opioid cytokines with an apparent molecular mass of 15 kDa and 67 kDa that are recognized both by anti-interleukin-4 and anti-dynorphin B antibodies. These putative new cytokines were named melatonin-induced-opioid (MIO). The most active and naltrexone-sensitive MIO was the smaller molecule, which was called MIO15 and found to act on an opioid-binding site present in adherent bone marrow cells. However, the hematopoietic action of MIO15 was dependent on the presence of colony-stimulating factors (CSF). To investigate this point, we studied the ability of melatonin to rescue granulocyte/macrophage colony-forming units (GM-CFU) in the bone marrow of tumor-free animals treated with cancer chemotherapeutic compounds. We found that melatonin not only is unable to protect bone marrow GM-CFU unless the mice are transplanted with Lewis lung carcinoma (LLC), but also that melatonin seems to increase the myelotoxicity of cyclophosphamide in tumor-free mice. In both tumor-bearing or healthy mice, the effect of melatonin is negated by naltrexone, indicating the involvement of MIO15. Competition studies classified the target opioid-binding site as a kappa-opioid receptor with low affinity in tumor-free mice and high affinity in LLC-implanted mice. LLC is known to release CSF. Consistently, addition of CSF in the form of lung-conditioned medium (LCM) to adherent bone marrow cells increased the affinity of the kappa-opioid receptor. Addition of antigranulocyte/macrophage colony-stimulating factor (GM-CSF) mAbs neutralized the effect of LCM. In conclusion, the affinity state of the kappa-opioid receptors in stromal bone marrow cells seems to modulate the hematopoietic effect of melatonin and/or MIO15.