ERK2 Substrate Binding Domains Play Distinct Roles in Megakaryocytic-Erythroid Lineage Progression and Mediates Clonal Fitness in Myeloproliferative Neoplasms

The RAS/MAPK signaling cascade is a critical mediator of multi-lineage blood cell development and function. Its terminal signaling component, ERK2, is part of a family of serine/threonine kinases that phosphorylates a variety of substrates that regulate cellular differentiation and survival. These substrates can be defined by their capacity to interact with the common docking site (D-domain) and the DEF binding pocket (DBP domain) of ERK2. We have previously demonstrated that disabling the DBP domain while preserving D-domain function accelerates the progression of a mouse JAK2-V617F driven model of polycythemia vera toward myelofibrosis (Zhang et al, Blood, 2022. PMID: 35436326). Further investigation is warranted to appreciate how ERK2 binding domains play distinct roles in the development of myeloproliferative neoplasms (MPN).

To investigate the functional role of the D-domain in vivo, we disabled D-domain function by utilizing CRISPR-Cas9 to knock in a FLip-EXcision switch into exon 7 of ERK2, where an inverted exon 7 containing a point mutation (D319N) is flipped into an open reading frame upon expression of Cre recombinase. Through this method, we were able to avoid embryonic lethality associated with inactivating the D-domain and study its effects on hematopoiesis by conditionally activating ERK2^D319N via excision of the ERK2^Fl allele in hemizygous ERK2^D319N/Fl mice under control of the Vav promoter. Together with the ERK2^Fl/Y261A DBP mutant mouse, we are the first to generate a complement model of ERK2 domain specific function in the context of hematological diseases.

Expression of the D mutated allele was exclusively restricted to lineage depleted cells where it is completely absent in cells derived from peripheral blood, spleen, and thymus of the mice. Flow cytometric analysis of ERK2^Fl/+(WT), ERK2^Fl/Fl (KO), ERK2^Fl/Y261A(DBP) and ERK2^Fl/D319N(D) mice reveals expansion of LSK (Lin^-Sca1⁺cKit⁺) cells in the DBP mutant mice which is not observed in the D mutant mice. This expansion of LSK cells is attributed by an accumulation of LT-HSC/MPP1 (LSK CD48^-CD150^-) indicating an early differentiation block in hematopoiesis when the DBP is inactivated. Interestingly, there is a significant increase in megakaryocyte restricted CD41 expressing LSK cells in the DBP mutant mice but not in the D mutant mice, suggesting a D-domain specific role in megakaryocyte development.

To delineate the population dynamics within the hematopoietic compartment, we performed high throughput scRNA-seq and CITE-seq on lineage depleted cells from WT, KO, DBP, D mutant mice to generate single cell transcriptomes at spatiotemporal resolution. Lineage depleted cells from the DBP mutant mice but not the D mutant mice are characterized by the expansion of megakaryocytic-erythroid progenitors and increased expression of canonical megakaryocyte markers within the LSK compartment. These results implicate that the D domain may serve to mobilize megakaryocytes to initiate myelofibrosis in the presence of an oncogenic driver e.g., JAK2^V617F, MPL^W515L/K, and CALR.

Indeed, we have shown in an adoptive transfer model of JAK2^V617F expressing DBP mutant progenitor stem cells that there was significant megakaryocyte hyperplasia in the bone marrow of recipient mice. Retroviral expression of MPN driver mutants JAK2^V617F and CALR in DBP mutant hematopoietic progenitors enhanced their clonogenic potential and serial replating capacity in contrast to the D mutant counterparts. This supports our hypothesis that the D domain maintains MPN clonal fitness and arbitrates megakaryocyte dysfunction in myelofibrosis. Finally, we sought to explore the therapeutic avenue of combinatorial drug synergies using D-domain inhibitors, Compound 76, 30g and 30h and JAK1/2 inhibitor, Ruxolitnib. Sub-efficacious dose of Ruxolitinib (<100nM) in combination with compound 30g significantly inhibited cell proliferation and survival of human MPN cell lines in contrast to treating with Ruxolitinib alone. Together, our findings support the ERK2-D domain as a suitable target for the treatment of myeloproliferative neoplasms.

No relevant conflicts of interest to declare.