In adult mice suffering from a phenylhydrazine (PHZ)-induced hemolytic anemia, erythropoietic islands were observed in the liver. These islands were studied with the light and electron microscope. Within two days after the beginning of four daily injections of PHZ, erythroid elements appeared in the sinusoids and central veins. A maximum number of erythroblasts was found on day 7. Light and electron microscopic observations revealed that the erythropoietic islands consisted of centrally located macrophages (CM) with a Kupffer cell-like morphology, surrounded by erythroblasts, which were often of the same maturation stage. CM in central veins (CM-V) and in sinusoids (CM-S) were found to have a different morphology. The CM-V phagocytized less circulating red blood cells and were in contact with a smaller number of erythroblasts. Furthermore, the contact areas between erythroblasts and CM-S extended for a much longer distance than those between erythroblasts and CM-V. The progenitor cell for the CM-V is most likely a monocyte, since cells which were morphologically determined as monocytes were found to appear on the first day of the PHZ treatment and differentiated into macrophages within about 2 days. The origin of the CM-S population was less clear, but could be monocytic as well. These data are tentatively explained as a migration of a progenitor of a cellular component of the erythroid microenvironment into the liver after appropriate stimuli. In contrast to fetal liver erythropoiesis, erythroblasts in the adult liver occurred only incidentally extrasinusoidally. Furthermore, specialized membrane contacts between erythroblasts and CM or hepatocytes could not be observed in adult liver. Ferritin could not be detected on the erythroid cell membrane or located in coated vesicles. Also, no ferritin could be observed within or attached to the finger-like processes of CM. The observations suggest that the coated vesicles in erythroid elements are partly exocytotic vesicles and are not specific for ferritin transport. The morphological aspects of PHZ-induced extramedullary erythropoiesis is discussed in relation to the hemopoietic microenvironment.

Furthermore, the authors are grateful to Mr. J.G.H. Fengler and Mr. T. van Os for printing the photographs and to Mrs. T. Schreuders-Henderson and Miss C. Clerkx for typing the manuscript