Modeling Calreticulin-Mutant Myeloproliferative Neoplasms with Isogenic Induced Pluripotent Stem Cells

Myeloproliferative neoplasms (MPN) are characterized by the excessive production of one or more myeloid lineages and a propensity to progress to acute leukemia. In 2013, mutations in the CALR gene, encoding calreticulin, were identified in patients with MPN, mutually exclusive to the previously identified JAK2 and MPL (TPO-R) mutations. CALR mutations are frameshift mutations - typically a 52-bp deletion (type 1) or a 5-bp insertion (type 2) - that result in a novel C-terminus. The discovery of mutations in a ubiquitously expressed multifunctional protein like calreticulin was unanticipated. Subsequent studies found that CALR mutations lead to activation of JAK/STAT, mediated through aberrant interactions between mutant CALR and MPL, thus presenting an excellent opportunity for targeted therapy. However, the mechanism of MPL activation remains largely unexplained with prior studies using cell lines with exogenous expression of CALR and MPL following transfection.

To create a more physiological cellular model to study the effects of CALR mutations, we established multiple iPSC lines from two patients with CALR-mutant MPN - one type 1-like (del34) and one type 2 (ins5) -, as well as from one patient with JAK2V617F MPN. All iPSC lines were confirmed to harbour the CALR or JAK2V617F mutation found in the corresponding patient, to express mutant calreticulin, as detected by flow cytometry using an antibody which specifically recognizes the novel calreticulin C-terminus, and to be karyotypically normal. Genetically matched iPSC lines with WT JAK2 could also be generated from the JAK2V617F (but not the CALR-mutant) patient cells in the same reprogramming round. CRISPR gene editing was used to generate isogenic CALR-corrected lines from both CALR-mutant patients. Furthermore, in order to facilitate biochemical studies, we used CRISPR to introduce a V5 epitope tag in one allele of the endogenous mutant or WT CALR gene, in mutant and isogenic corrected iPSC lines, respectively. We optimized an in vitro differentiation protocol for efficient derivation of megakaryocyte (MK) progenitors from iPSCs and found disease-relevant phenotypes, mainly TPO-independent MK colony formation in semi-solid media, which is the phenotypic hallmark of ex vivo primary MPN cells. In the absence of TPO, JAK2 V617F, CALR-mutant type 1-like and CALR-mutant type 2 iPSCs generated 52.1%, 58.7±22.2% and 59.8±3.6%, respectively, of the number of MK colonies generated in the presence of TPO, as opposed to 10%, 8.8±1.8% and 0.5±0.9%, respectively, for the matched WT JAK2, the corrected CALR-mutant type 1-like and the corrected CALR-mutant type 2 iPSCs. Isolated CALR mutant iPSC-derived CD41a⁺ MK progenitors had increased phosphorylation of STAT5 following cytokine starvation as compared to isogenic corrected and non-isogenic normal cells. CALR-mutant cells expressed equal transcript levels of the WT and mutant CALR alleles. However, mutant CALR protein levels were severely reduced, at levels 1~12% of those of the WT protein. This is consistent with previous studies documenting instability of mutant calreticulin. Transcriptomics (RNA-seq) and proteomics analyses of CD41a⁺-sorted MK progenitors derived from CALR mutant and isogenic corrected iPSCs are ongoing.

These iPSC models offer the opportunity to study the effects of CALR mutations in a cellular context with both MPL and CALR (WT or mutant) expressed from their endogenous loci. They thus provide a powerful platform to investigate the disease mechanisms underlying CALR-mutant MPNs and to perform small molecule and genetic (CRISPR) screens to identify new therapeutic targets.