The discovery that several ribosomal protein genes can be mutated in DBA suggests that ribosomal protein gene mutations may account for many or all cases of DBA, and focuses attention on the ribosome. While experiments in yeast and recently in mammalian cells show that RPS19 depletion or mutation leads to a block in ribosomal RNA biosynthesis, this result does not explain why erythropoiesis is so severely affected in DBA. We hypothesize that during fetal development immature erythroid cells proliferate more rapidly than other lineages and therefore require very high ribosome synthetic rates to generate sufficient capacity for translation of erythroid specific transcripts that must take place before these unique cells enucleate; furthermore, we postulate that a block in ribosome biogenesis or reduced protein synthetic capacity that occurs in mutant DBA cells leads to loss of proliferation and cell death of rapidly dividing cells, but survival and normal differentiation of cells that are dividing more slowly, yielding fewer (macrocytic) erythrocytes. To test this kinetic hypothesis we infected primary mouse fetal liver cells with siRNAs to RPS19 and compared proliferation, differentiation, RNA biogenesis and cell cycle status in wild type and knockdown cells. Mouse fetal liver cells were double-labeled for erythroid-specific TER119 and non erythroid-specific transferrin receptor (CD71) and analyzed by flow-cytometry. E14.5 fetal livers contain at least five distinct populations of cells, defined by their characteristic staining patterns. We purified the most primitive progenitor cells by depletion of mature TER119high cells. During a two-day period the number of erythroblasts increases 15-30 fold, corresponding to 4–5 cell divisions, which correlates well with the number of terminal cell divisions that a CFU-E goes through to generate terminally differentiated erythrocytes. The progenitor cells divide twice during the first 24 hours in erythropoietin (EPO); during the next 24 hours on fibronectin but no EPO, differentiated cells are produced in another 2–3 divisions. The retrovirus infected siRNA RPS19 knockdown cells show reduced proliferation of FACS sorted GFP positive cells at 48 hours. Although the cell yield is reduced, the differentiation pattern of the surviving GFP positive cells is similar to that of the controls. We next measured RNA content of wild type cells at 0, 24 and 48 hours. During the first 24 hours cell number increases 3–4 fold; remarkably, there is a 6-fold increase in RNA content during the same period, suggesting that the cells accumulate an excess of ribosomal RNA (80% of measured RNA) during early erythropoiesis. This was confirmed by quantitative real time PCR of rRNA. From 24–48 hours the cells decrease in size as they mature, and RNA yield per cell decreases; however, cell number increases markedly so the net effect is that total RNA in the culture plateaus or decreases. Because the siRNAs are not expressed until 24–48 hours, we modified the culture system to allow expansion without differentiation of immature cells in EPO, IGF-1 and dexamethasone. Under these conditions proliferation in siRNA expressing precursors is reduced. Cell cycle analysis shows a reduced proportion of cells in G1 or S phase and an increase in G2/M in the knockdown cells. Taken together, these data suggest that RPS19 insufficient erythroid cells proliferate poorly because of inadequate accumulation of ribosome synthetic capacity. The surviving cells differentiate normally but slowly, giving rise to macrocytes. In conclusion, kinetic considerations can explain the erythroid deficiency in DBA.

Disclosure: No relevant conflicts of interest to declare.