Disease Phenotype and Clonal Amplification in Calreticulin del52 and ins5 Knock-in Mice Are Dependent on the Type of Mutations and Gene Dosage

Introduction

BCR-ABL-negative myeloproliferative neoplasms (MPNs) result from the transformation of a hematopoietic stem cell (HSC). Somatic mutations in the calreticulin (CALR) gene are associated with approximately 30% of essential thrombocythemia (ET) and primary myelofibrosis (PMF). All CALR mutations induce a frameshift to the same alternative reading frame generating a new C-terminal tail. The two most frequent CALR mutations are a 52 bp deletion (del52) or type 1 and a 5 bp insertion (ins5) or type 2. In patients, del52 and ins5 are equally found in ET but del52 is more frequent in PMF. In mouse retroviral model, del52 mice progress from ET to myelofibrosis (MF) while ins5 mice remain mostly with an ET.

Methods

In order to study the effect of endogenous levels of del52 and ins5 in hematopoiesis, we generated conditional knock-in (KI) mice expressing the murine CALR del52 or ins5 with the human mutated C-terminal tail under the control of a Scl-driven tamoxifen-inducible Cre recombinase (Scl-CreER^T). We have also used Ubi-GFP transgenic mice to perform competitive engraftments.

Results

After tamoxifen-induction, both del52 and ins5 KI mice developed a rapid thrombocytosis, more severe in the homozygous than the heterozygous setting. In contrast, leukocytosis was observed only in homozygous setting. At similar zygosity, del52 induced a higher thrombocytosis compared to ins5. After 10 months of induction, both the bone marrow (BM) and the spleen of homozygous del52 KI mice and, to a much lower extent of homozygous ins5 KI mice, presented a significant increase in megakaryocytes (MKs) and in MK progenitors by flow cytometry. Von Willebrand factor staining showed that both del52 and ins5 homozygous mice displayed giant polylobulated MKs, associated with a similar increase in ploidy (mean ploidy 32N-33N). Heterozygous del52 presented also an increase ploidy of MK (mean of 25N) compared to controls (mean of 17N), whereas the MK ploidy of heterozygous ins5 mice was similar to control mice. The increase in number and size of MKs in homozygous del52 mice partially explained the significant decrease in BM cellularity and the splenomegaly. Moreover, we observed a decrease in BM erythroblasts and, in spleen, an increase in both erythroblast and granulocytic precursors together with a decrease in lymphocytes associated with a major disorganization of white pulp territories. Thus, the del52 homozygous KI mice developed features of a MF-like disease further illustrated by the presence of reticulin fibers stained with silver, mainly in spleen. Presence of fibrosis was not as pronounced in heterozygous del52 mice and more rarely observed in spleen of homozygous ins5 KI mice.

In homozygous del52 KI mice, there was a significant amplification of the HSC compartment in both BM and spleen that was stronger than in homozygous ins5 mice. To study whether del52 and ins5 could provide a competitive advantage to HSCs, we performed BM transplantation with increasing percentages of non-induced homozygous del52 or ins5 cells with wild-type GFP⁺ cells into lethally-irradiated recipient mice. The homozygous del52 BM cells strongly competed wild-type hematopoiesis from an initial engraftment as low as 10% of mutated clones, reaching 100% in both blood myeloid cells and BM HSC compartments at 4 months. In contrast, out-competition of wild-type hematopoiesis by homozygous ins5 cells was slower, especially when less than 50% of mutated cells were initially engrafted suggesting that del52 provides a stronger advantage to the HSCs than ins5.

Conclusion

In conclusion, these results demonstrate that modeling CALRdel52 and ins5 mutations in mice can successfully recapitulate the differences in phenotype observed in patients, i.e del52 KI mice recapitulate an ET progressing to MF while ins5 KI mice only mimic an ET. This might be explained by a more profound effect of del52 than ins5 at the level of HSC. These KI mice offer solid in vivo models to investigate the mechanism of action of both types of mutations on HSCs and MKs and will be used to test new therapeutic approaches.

Disclosures: No relevant conflicts of interest to declare.