Calmodulin Inhibition Rescues DBA Models with Ribosomal Protein Deficiency through Reduction of RSK Signaling

Ribosomal protein (RP) mutations are found in many diseases, including Diamond Blackfan anemia (DBA), where defective erythropoiesis, craniofacial abnormalities and increased cancer risk are major complications. RP mutations are thought to cause p53 activation through accumulation of free RPs that bind and sequester MDM2, the negative regulator of p53. We previously characterized a zebrafish mutant in rps29, a ribosomal gene found mutated in DBA patients. Rps29^-/- embryos have hematopoietic and endothelial defects, including decreased cmyb and flk1 expression and defects in hemoglobinization. Consistent with other animal models of RP dysfunction, p53 knockdown in rps29^-/- embryos rescued these defects. To uncover novel compounds that correct the phenotypes of DBA, we performed a chemical screen in rps29^-/- embryos. Several structurally distinct calmodulin (CaM) inhibitors successfully rescued hemoglobin (Hb) levels in the mutant embryo. To confirm that CaM inhibitors could rescue mammalian models of DBA, we tested them in human and murine models. Treating cord blood-derived CD34+ cells deficient in RPS19, the CaM inhibitor trifluoperazine (TFP) relieved the erythroid differentiation block. Injection of TFP in a DBA murine model significantly increased red blood cell number and Hb levels and reduced p53 activity in the bone marrow. Of note, the effect of TFP was specific to RP deficiency and had no effect on erythroid differentiation or p53 activity in WT cells or mice. This prompted us to hypothesize that TFP is selectively blocking signaling in the RP deficient state, possibly through inhibition of CaM-dependent kinases. In vitro kinase profiling of over 100 kinases revealed that TFP and other positive hits from our screen inhibited the activity of p70 ribosomal S6 kinase (p70S6K) and multiple members of p90 ribosomal S6 kinase (RSK) family. RSKs are highly conserved serine/threonine kinases that regulate cell growth, migration, and survival. They can activate mTOR and directly phosphorylate RPS6. Strikingly, RSK protein levels are elevated in lysates from rps29^-/- embryos and treatment with RSK or p70S6K inhibitors increased Hb in rps29^-/- embryos in vivo, mimicking CaM inhibitors. RSK phosphorylation is also increased in human RPS19-deficient peripheral blood-derived CD34+ cells compared to WT cells. Treatment with TFP reduced RSK phosphorylation and decreased signaling downstream of RSK, including p70S6K, but only in the presence of RP deficiency. Similarly, TFP reduced p53 and phosphorylation of p53 at S392, but only in the presence of RP deficiency. An in vitro kinase assay determined that p70S6K directly phosphorylated p53 at S392 but not other commonly modified residues, S15 and S20. Mass spectrometry analysis of posttranslational modifications of p53 isolated from TFP treated cells revealed a significant reduction of peptides that contain phosphorylated S392. We hypothesized that TFP inhibited phosphorylation of S392. To test this hypothesis, phosphomimetic mutants were transfected into Saos2 cells and p53 transcriptional activity was evaluated using p21mRNA levels. TFP treatment of cells containing WT p53 or a negative control transactivation domain mutant, p53-S15D, resulted in a 4-fold reduction in p21 mRNA levels, while cells containing p53-S392D had no reduction in p21 mRNA in response to TFP. In conclusion, we have shown that RP deficiency increases RSK phosphorylation and CaM inhibitors decrease signaling downstream of RSK, which leads to a reduction of p53 activity and rescues the phenotypes of multiple in vitro and in vivo models of DBA. Our data strongly suggests that CaM inhibitors may be effective therapies for DBA patients, and a clinical trial is being planned with TFP for 2017.