Knock-Ins of Type-2 Calr Mutants Cause Myeloproliferative Neoplasm (MPN)-like Hematopoiesis in Mice

MPNs, including polycythemia vera, essential thrombocythemia (ET) and myelofibrosis (MF), are characterized by proliferation of mature myeloid cells. A somatic mutation in a hematopoietic stem cell (HSC) that activates JAK/STAT pathway drives MPN. Following the JAK2^V617F, the CALR insertion/deletion mutations (indels) are the second most frequent driver and present in 20-30% ET and primary MF. Two major indels, a 52-bp deletion (type 1, p.L367fs*46) and a 5-bp insertion (type 2, p.K385fs*47), account for 80% of the CALR mutations. In addition, there have been more than 100 other indels, which can be classified as type 1- and type 2-like mutations based on the length of negatively-charged amino acid (AA) stretch at the C-terminal side of mutated CALR. Patients with type 1/type 1-like CALR mutations exhibit more incidence of MF while type 2/type 2-like mutations are associated with higher platelet counts in ET (Petra et al, Leukemia, 2016). In previous studies, mice carrying type 1/type 1-like mutations, including knock-in (KI) models, showed mild ET- or MF-like hematopoiesis. Although wild-type (WT) CALR AA sequences are highly conserved between human and mouse, there have been no KI models of type 2/type 2-like CALR mutations. Here, we generated 2 lines of KI mice carrying type 2-like Calr mutations, 2-bp insertion (CR2i, p.K378fs*53) and 10-bp deletion (CR10d, p.K375fs*52), using the CRISPR/Cas9 method. Both KIs removed KDEL, altered AA charges and increased values for isoelectric point, which are similar to type 2/type 2-like mutations in MPN patients. Compared with WT mice, peripheral platelets (1277 ± 228 vs 1560 ± 344 x 10⁹/L, p = 0.004) and leukocytes (14.4 ± 3.7 vs 18.7 ± 4.9 x 10⁹/L, p = 0.006) were increased in CR10d mice, whereas blood cell counts were not different between CR2i and WT mice. In FACS, both CR10d (p = 0.04) and CR2i (p = 0.04) mice exhibited an increased myeloid-cell ratio in bone marrow (BM). Splenomegaly was not present, but histopathological study showed a significant increase and accumulation of large megakaryocytes in BM and spleen of both KI mouse lines. BM fibrosis was not present in any sample. Therefore, CR10d and CR2i mice mimicked ET-like and unclassifiable MPN-like hematopoiesis, respectively. Next, we studied the basis of MPN-like hematopoiesis in CR10d and CR2i mice. Colony forming-cell assay in the presence of cytokines showed reduced growth of CFU-E, especially in CR2i mice (p = 0.01) compared with WT mice, while there was no difference in growth of CFU-Mk between CR10d or CR2i mice and WT mice. TPO-independent colony growth was not observed in both KI mice. Correspondingly, FACS showed comparable expression of phospho-STAT3 (pSTAT3) in BM cells between CR10d or CR2i mice and WT mice in the absence of TPO. However, pSTAT3 was significantly upregulated both in CR10d and CR2i mice compared with WT mice in the presence of TPO, suggesting that high sensitivity of HSCs or progenitor cells to TPO contributes to MPN phenotype in these mice. Thus, we investigated HSC function by a competitive repopulation assay, in which we transplanted a mixture of BM cells from KI mice (Ly5.2) and Ly5.1 mice at a 1:1 ratio into lethally irradiated Ly5.1 mice, showed reduced repopulating capacity, especially in CR2i mice. In the second transplant recipients, cells derived from either CR2i or CR10d mice were markedly diminished, suggesting the reduced self-renewal capacity of an HSC carrying a type 2/type 2-like Calr mutation. Finally, we performed RNAseq for FACS-sorted HSC-enriched lineage^-Sca1⁺Kit⁺ (LSK) cells, which revealed that approximately 70% of genes among differentially expressed genes were commonly upregulated or downregulated in CR2i and CR10d mice, suggesting a similarity in gene expression profile of LSK cells of these KI mouse lines. As a result, there were several pathways commonly affected in both CR2i and CR10d mice in gene set enrichment analysis, including upregulation of JAK/STAT pathway (FDRq = 0.060 in CR2i and 0.111 in CR10d). On the other hand, targets of polycomb recessive complex 2, which are important for HSC functions in MPNs (Ueda et al, Blood Adv, 2017), were downregulated in both KI mouse lines (FDRq = 0 in both CR2i and CR10d), possibly explaining the reduced repopulating capacities of CR2i and CR10d HSCs. In conclusion, our data indicate that type 2/type 2-like Calr mutation can cause MPN-like hematopoiesis. For disease progression, further mechanism may be required.