Negative Autoregulation by Fas Stabilizes the Erythroid Progenitor Pool and Accelerates the Erythropoietic Stress Response

Abstract 2045

Signaling and transcriptional networks frequently contain negative autoregulatory feedback loops, where gene products negatively regulate their own induction or activation. These negative autoregulatory motifs are predicted to exert dual functions, accelerating gene induction, and providing stable gene expression levels in the face of the random perturbations inherent to biological systems. These predictions were confirmed experimentally in synthetic transcriptional circuits [1,2], but it is unknown whether they also hold in naturally occurring higher level biological networks. Here we studied the role of negative autoregulation by erythroid progenitors in the control of erythropoiesis.

Erythropoietic rate, which may increase ten fold its basal rate during hypoxic stress, is dependent on the size of the erythroid progenitor pool, in turn regulated by the hormone erythropoietin (Epo). We recently found that, in addition, early erythroblasts negatively regulate their own numbers, through their co-expression of the death receptor Fas and its ligand FasL. Here we investigated the role of this negative autoregulation using Fas or FasL-deficient mice. We used the naturally-occuring mutant mouse strains, lpr and gld, deficient in Fas and FasL, respectively, back crossed onto the Rag1^-/- mutant background, in order to avoid the autoimmune syndrome associated with Fas mutation. We proceeded to examine basal and stress erythropoiesis in the gld-Rag1^-/- and lpr-Rag1^-/- mice, and in matched Rag1^-/- controls.

We found that, in the basal steady state, the average size of the spleen early erythroid progenitor pool in gld-Rag1^-/- and lpr-Rag1^-/- mice increased 1.5 to 2 fold, consistent with loss of a negative regulator. Further, gld-Rag1^-/- mice had a significantly elevated hematocrit in spite of normal Epo blood levels. The hematocrit was normal in the lpr-Rag1^-/- mice, but Epo levels in this strain were significantly lower than normal. Taken together, these genetic mouse models show that Fas-mediated apoptosis of early erythroblasts in spleen negatively regulates erythropoietic rate in the basal state.

We also found that the size of the progenitor pool was highly variable between individual Fas-deficient mice, suggesting reduced ability to maintain a stable steady-state erythorpoietic rate. In addition, gld-Rag1^-/- and lpr-Rag1^-/- mice had a significantly delayed erythropoietic stress response. Following an injection of a single dose of Epo (300 U/25 g), the early erythroblast population in spleen, ‘EryA’ (Ter119^highCD71^highFSC^high, [3]) expanded 30 to 60 fold its basal size. However, this expansion was significantly delayed in gld-Rag1^-/- and lpr-Rag1^-/- mice. Specifically, on day 2 of the stress response, control Rag1^-/- mice had a 30% larger EryA progenitor pool compared with lpr-Rag1^-/- mice, a difference equivalent to 10 fold the size of the basal EryA pool. Consequently, control mice achieved a higher hematocrit 24 hours earlier than mutant gld-Rag1^-/- and lpr-Rag1^-/- mice.

We propose that the larger expansion of EryA cells during the stress response in control mice is due to the recruitment of a reserve population of Fas-positive EryA. This reserve population, absent in mice deficient in the Fas pathway, undergoes Fas-mediated apoptosis in the basal steady state. However, high Epo levels during the stress response suppress Fas expression [3], rescuing these cells from apoptosis and accelerating the stress response.

These findings show, using genetic mouse models, that the stability of stead-state erythropoietic rate and its rapid stress response are key outcomes of negative autoregulation within the erythroid progenitor pool. Furthermore, they show experimentally that dynamic properties of negative autoregulatory loops in simple low-level networks are also exerted in the context of complex inter-cellular, tissue level networks such as those that regulate erythroipoietic rate.