Deletion of Calr By CRISPR/Cas9 Genome-Editing Mimics Abnormalities of Megakaryocytopoiesis Induced By Calr Del52 Mutation

INTRODUCTION. Calreticulin (CALR) is mutated in 20-25% of patients with myeloproliferative neoplasms (MPN). Most common abnormalities are deletion of 52bp (T1) and insertion of 5bp (T2) in exon 9, causing recurrent frameshift and generation of novel C-terminal domain. Interactions of mutant CALR with the thrombopoietin receptor MPL contribute to abnormal megakaryocytopoiesis of MPN (Chachoua I, Blood 2016; Araki M, Blood 2016; Elf S, Cancer Discov 2016), but pathogenic mechanisms remain largely unknown.

METHODS. Rabbit antibodies (pAb) against mutated CALR were obtained and extensively characterized for their binding to the mutated CALR only. CALR knock-out (KO) and CALR-T1 variants were generated by CRISPR/Cas9 technology in K562 cells; transient expression of CALR-WT, T1 and T2 was obtained by transfecting KO cells. UT7 cells stably expressing CALR-T1, T2 and WT were generous gift of W. Vainchenker. For in-vivo experiments, nude mice were injected subcutaneously with K562 CALR-KO, WT, T1 and T2 cells and tumor mass was measured.

RESULTS. We found that genome-edited K562 CALR-T1 cells expressed similar levels of WT and del52 CALR transcripts (by qRT-PCR) while no amplification was obtained in KO cells; correctly edited sequence in T1 cells was assessed by Sanger. Presence of both WT and T1 CALR protein with the predicted molecular weight was confirmed by immunoblotting; no immunoreactivity was found in KO lysates. Interestingly, we identified a cleaved CALR product (about 40KDa) in K562 CALR-T1 and UT7 CALR-T1/T2 cells, whose generation was inhibited by 15mM ROS-inhibitor N-acethyl-L-cysteine (NAC), suggesting that the mutated protein is involved in oxidative stress pathway. By confocal microscopy with mutation-specific antibodies, the mutant CALR was localized in both cytosolic compartment and cell membrane of K562 CALR-T1 and UT7 CALR-T1 and -T2 cells. No significant differences in cell cycle and proliferative rate of K562 KO and T1 cells were observed in comparison to parental cells. However, K562 KO and T1 cells were more resistant to apoptosis compared with parental cells (p<0.05). When induced to Mk commitment in liquid cultures with phorbol-myristate-acetate (PMA), KO and T1 cells both showed accelerated and enhanced expression of CD41/CD61 vs parental cells (2.3- and 3-fold higher, respectively; p<0.01). These data were further confirmed in K562 KO and KO cells overexpressing CALR-WT, T1 and T2, where the proportion of CD41/CD61 positive cells raised from 30% in WT cells to 55%, 62%, 70% respectively in T1, T2 and KO cells (p<0.01). We also produced CRISPR/Cas9 CALR-KO CD34+ from cord blood; clonogenic assays showed that CALR-KO CD34+ cells generated CFU-Mk and BFU-E colonies 8- and 2.5-fold more, respectively, than untouched cells (n=5; p<0.05). Overall, these data indicated that deletion of CALR by genome editing resulted in promotion of megakaryocytopoiesis, and to less extent of erythropoieisis, in normal CD34+ cells, mimicking the effects obtained in K562 CALR-KO and T1 induced to Mk differentiation by PMA.

A number of differential expressed mRNAs was found by target profiling in K562 KO and T1 cells. One recurrently affected gene category was "integrin-binding and adhesion molecules"; in particular, ICAM-1, Osteonectin and Stathmin, involved in integrin and collagen binding and cytoskeleton regulation, were significant down-regulated (-40%, -50%, -60% respectively, vs parental cells; p<0.05). Clarification of their role in normal megakaryocytopoiesis is ongoing. The functional relevance of those deregulated genes was supported by in-vivo experiments, where K562 CALR-WT, KO or T1 cells were injected subcutaneously in different flanks of nude mice. We found that K562 KO and T1 cells were both unable to generate a measurable tumor mass while WT cells rapidly produced visible tumors up to 2g weight 20 days post-injection, suggesting defective anchorage-mediated tumor growth. These findings are in line with reports indicating inability of CALR-KO solid tumor cells to engraft (Feng R, Diagn Pathol, 2015; Feng S, Clin Cancer Res, 2014).

CONCLUSIONS. These results suggest that loss of wild-type CALR by CRISPR/Cas9-mediated target deletion in cell lines and primary CD34+ cells phenocopies the effects of CALR del52 mutations on megakaryocytopoiesis.