To date no hematopoietic progenitors of dendritic Langerhans' cells (DLC), which represent an highly efficient class of antigen presenting cells, have been identified or the cytokines they elaborate have been defined. Here we describe an acute leukemia patient whose blasts (90-96% in peripheral blood and bone marrow) had a phenotype consistent with putative progenitors of DLC. The patient was treated with ara-C and VP-16 but did not achieve remission. The blasts had lobulated nuclei, no cytoplasmic vacuolation or Auer rods and were weakly positive for acid phosphatase and non-specific esterase and negative for PAS, granzyme A, dipeptidyl aminopeptidase IV, ATPase/ADPase and lysozyme production. The blasts were positive for CD1a, CD4, CD16, CD35, HLADR, HLADQ, CD11b, CD11c, CD14, CD33, CD34, CD11a, CD71, CD19, CD25, IL-2R beta and negative for CD2, CD7, CD8, CD10, CD22, CD56, CD57, surface or cytoplasmic CD3, TCR delta and TCR beta, HTLV-1p19 and P-glycoprotein. On liquid culture with or without 5 x 10(-9) M 12-O-tetradecanoylphorbol-13-acetate (TPA) for 3 days, the blasts formed aggregates of proliferating and elongating cells on the wall of the flasks with a decline in CD34, numerous dendritic processes appeared on the cells and there was strong positivity for ATPase/ADPase, but no other changes in phenotype. No macrophages were observed, indicating derivation from separate DLCs. Cytogenetic analysis showed chromosomal abnormalities and electron microscopy showed Birbeck granules. Southern blotting of DNA showed rearrangement of one allele for both JH and TCR beta but no HTLV-1 related sequences. Culture supernatants from blasts cultured with or without TPA showed the production of large amounts of IL-8, IL-6, TNF-alpha, MIP-1 alpha, IL-10 and interferon gamma and modest amounts of IL-1 alpha, GM-CSF and stem cell factor. The presence not only of CD1a, HLADR, HLADQ and many other characteristics including Birbeck granules, but also differentiation along the lines of DLC with appearance of dendritic processes on the cells and expression of ATPase/ADPase activity, indicate that the leukemic blasts in our patient represented a leukemic counterpart of normal progenitors of DLC and the leukemia a new entity which could possibly be classified as AML-M8. Lastly, many pro-inflammatory cytokines produced by DLC could contribute to inflammation and IL-10 to immunosuppression.