Physical Interaction Between Mutant Calreticulin and the Thrombopoietin Receptor Is Required for Hematopoietic Transformation

Somatic mutations in calreticulin (CALR), an endoplasmic reticulum (ER) chaperone protein, are found in up to 40% of patients with myeloproliferative neoplasms (MPN). All pathologic CALR mutations are out-of-frame insertion and/or deletions (indels) in exon 9, generating a 1 base-pair (bp) frame shift and a common mutant-specific C-terminus, with the most common mutation being a 52 bp deletion (del52).

The observation that CALR mutations are mutually exclusive with other MPN-initiating mutations such as JAK2V617F suggests a key pathogenic role for mutant CALR. To determine if mutant CALR alone is sufficient to induce MPN we began by over-expressing CALR-del52 in a retroviral bone marrow transplant (BMT) mouse model. We found that CALR-del52-expressing mice develop thrombocytosis and megakaryocytic hyperplasia, recapitulating the megakaryocyte-specific phenotype of CALR-mutant MPN patients. These findings suggest that the thrombopoietin receptor, MPL plays a key role in the pathogenesis of mutant CALR-driven MPN.

To evaluate the role of MPL in mutant CALR driven oncogenesis, we over-expressed CALR-del52 in interleukin-3 (IL-3)-dependent Ba/F3 hematopoietic cells. We found that CALR-del52 over-expression results in transformation to IL3-independent growth only in Ba/F3 cells co-expressing MPL, but not in parental Ba/F3 cells or Ba/F3 cells co-expressing the EPO receptor (EPOR) or the G-CSF receptor (GCSFR). We found similar results in human cytokine-dependent UT-7 cells. We also introduced +1 frameshift mutations into the endogenous Calr locus in Ba/F3-MPL cells using CRISPR/Cas9 gene editing and successfully engendered IL-3 independent growth, indicating that endogenous levels of mutant Calr expression are sufficient for transformation. Together, these data indicate that MPL is specifically required for the transforming capacity of mutant CALR.

Using RNA-sequencing followed by gene set enrichment analysis (GSEA), we confirmed that mutant CALR transformed Ba/F3-MPL cells display strong enrichment of Stat5 and Stat3 gene expression signatures. Concordantly, we also saw differential phosphorylation of Stat5 and Stat3 in these cells. Furthermore, we found that the IL-3 independent proliferation of mutant CALR expressing Ba/F3-MPL cells is decreased upon shRNA-mediated knockdown of Jak2, and that differential activation of Stat5 and Stat3 is abrogated by the JAK2 inhibitor, ruxolitinib. Together, these data demonstrate that mutant CALR signals through the JAK/STAT axis downstream of MPL.

We next sought to define the specific domains within mutant CALR required for oncogenic transformation. We found that neither expression of the mutant C-terminus alone nor expression of CALR lacking the C-terminus leads to cytokine-independent growth, suggesting that the novel C-terminus is necessary (but not sufficient) for transformation. We therefore generated an extensive series of truncation, domain deletion and point mutations within the C-terminus and assessed their respective transforming capabilities. Surprisingly, we found that the oncogenic activity of mutant CALR is not encoded within a specific sequence or domain of the mutant C-terminus. Rather, we found that the positive electrostatic charge of the mutant C-terminus is critical for its transforming capacity. Mutagenizing all 18 lysine/arginine residues (positively charged) within the C-terminus to a neutral glycine residue abrogates CALR-del52 transformation activity. In contrast, mutagenizing the 18 non-lysine/arginine residues within the C-terminus to glycine does not affect transforming activity, a remarkable finding considering that, in this mutant, 50% of the amino acids have been modified.

Finally, using co-immunoprecipitation assays we found that mutant CALR, but not wild-type CALR, physically interacts with MPL, and that neither the mutant C-terminus alone nor mutant CALR lacking the C-terminus can bind to MPL. This suggests that the tertiary structure of mutant CALR is required for binding to MPL. Moreover, we found that the ability of our engineered CALR mutants to bind MPL perfectly correlates with their ability to mediate transformation, suggesting that the interaction with MPL is critical for mutant CALR-mediated transformation.

Together, our findings elucidate a novel mechanism of pathogenesis in MPN and provide insights into how CALR mutations drive the development of MPN.