Harnessing a Novel Dyrk1a-Ablim2-MKL1 Regulatory Module in Megakaryocyte Morphogenesis to Enable Scalable Platelet "Pharming"

Growing clinical demands for platelet transfusions combined with supply limitations have created shortages which are trending toward a global crisis. Major efforts have been taken to address key issues of platelet sources, storage, and utilization. Recent progress in ex vivo culture-based production of megakaryocytes (Mk) and platelets, "pharming," has highlighted the potential for novel, donor-independent sources amenable to antigenic editing and cryo-stockpiling. Such cultures can be easily initiated from umbilical cord blood (CB) progenitors, induced pluripotent stem cells (iPSC), or directly re-programmed somatic cells. The major roadblock associated with these Mk sources consists of their fetal ontogenic status, which is beneficial for expansion but severely limits platelet production. The ability to elicit in pre-expanded Mk an adult program of morphogenesis (polyploidization, enlargement, and proplatelet formation) would enable circumvention of this scalability barrier.

A master regulator of adult Mk morphogenesis consists of the transcriptional coactivator MKL1 which undergoes nuclear translocation in response to RhoA-mediated actin polymerization, stimulated by thrombopoietin and environmental mechano-sensing. Nuclear MKL1 associates with the transcription factor SRF1 to upregulate cytoskeletal remodeling factors, including filamin A and Hic-5, that act as morphogenesis effectors. Our previous studies identified in infantile CB Mk a failure in MKL1 upregulation resulting from repression by the oncofetal RNA-binding factor IGF2BP3. Pharmacologic suppression of IGF2BP3 with BET inhibitors rescued MKL1 expression and improved platelet production but caused cycle arrest preventing polyploidization.

As an alternative approach to abrogate the fetal blockade in Mk morphogenesis, we sought to promote MKL1 activity by targeting a kinase, Dyrk1a, which had been shown to restrain MKL1 from nuclear translocation. Treatment of infantile CB Mk with a variety of Dyrk1-selective inhibitors including harmine and EHT 1610 strongly enhanced polyploidization (p = 0.015 and 0.009 respectively), enlargement (p < 0.005) , and in vitro platelet release (2 fold each, p = 0.001 and 0.007 respectively), attaining levels seen with adult Mk. When xenotransplanted into NSG mice, harmine-treated CB Mk demonstrated enhanced capability for in vivo platelet release (about 5 fold, p = 0.016). CB stem cells expanded with the AHR antagonist SR1 and an iPSC-Mk cell line also responded to Dyrk1 inhibition with robustly increased morphogenesis. Several findings implicated MKL1 in this response: 1) induction of nuclear translocation by the inhibitors, 2) induction of target genes (filamin A and Hic-5) by the inhibitors, and 3) loss of response to inhibitors in Mkl1-ko murine progenitors. Supporting Dyrk1a as a relevant target, mice with Mk-specific loss of one Dyrk1a allele (Dyrk1a^flox/wt;Pf4-Cre) displayed increases in platelet counts (p = 0.037) and marrow Mk polyploidization (p = 0.02). In addition, retroviral expression in human progenitors of a dominant negative Dyrk1a mutant K188R promoted Mk enlargement (p = 0.014). shRNA knockdowns could not be obtained due to toxicity of > ~60% loss of Dyrk1a.

To determine mechanisms for Dyrk1a control of morphogenesis, we analyzed the actin cytoskeleton, a key regulator of MKL1. Dyrk1 inhibition in all types of Mk progenitors (adult, infantile, and iPSC) induced assembly of cortical filamentous actin (F-actin), as detected by Alexa594-phalloidin staining. Prior studies showed cytoskeletal binding by Dyrk1a and direct phosphorylation of F-actin regulators N-WASP and Ablim1. A survey of human marrow expression patterns for candidate Dyrk1a substrates (The Human Protein Atlas) identified Ablim2, as showing a Mk-specific, cortical staining pattern. Dyrk1 inhibition increased Ablim2 levels ~5-fold in CB Mk (p < 0.005), and immunofluorescence displayed a cortical distribution similar to F-actin. Lentiviral shRNA knockdown of Ablim2 abrogated all effects of Dyrk1 inhibition, blocking: F-actin formation, MKL1 nuclear translocation, activation of the MKL1 targets, and Mk morphogenesis. These findings thus delineate a novel Dyrk1a-Ablim2-MKL1 regulatory module in Mk morphogenesis that can be manipulated to address the problem of scaling ex vivo production and might also serve as a future in vivo therapeutic target for thrombocytopenia.