Dihydroartemisinin inhibits the human erythroid cell differentiation by altering the cell cycle

Artemisinin derivatives such as dihydroartemisinin (DHA) induce significant depletion of early embryonic erythroblasts in animal models. We have reported previously that DHA specifically targets pro-erythroblasts and basophilic erythroblasts, when human CD34+ stem cells are differentiated toward the erythroid lineage, indicating that a window of susceptibility to artemisinins may exist also in human developmental erythropoiesis during pregnancy. To better investigate the toxicity of artemisinin derivatives, the structure–activity relationship was evaluated against the K562 leukaemia cell line, used as a model for differentiating early human erythroblasts. All artemisinins derivatives, except deoxyartemisinin, inhibited both spontaneous and induced erythroid differentiation, confirming that the peroxide bridge is responsible for the erythro-toxicity. On the contrary, cell growth was markedly reduced by DHA, artemisone and artesunate but not by artemisinin, 10-deoxoartemisinin or deoxy-artemisinin. The substituent at position C-10 is responsible only for the anti-proliferative effect, since 10-deoxoartemisinin did not reduce cell growth but arrested the differentiation of K562 cells. In particular, the results showed that DHA resulted the most potent and rapidly acting compound of the drug family, causing (i) the decreased expression of GpA surface receptors and the down regulation the γ-globin gene; (ii) the alteration of S phase of cell cycle and (iii) the induction of programmed cell death of early erythroblasts in a dose dependent manner within 24 h. In conclusion, these findings confirm that the active metabolite DHA is responsible for the erythro-toxicity of most of artemisinins used in therapy. Thus, as long as no further clinical data are available, current WHO recommendations of avoiding malaria treatment with artemisinins during the first trimester of pregnancy remain valid.