Mutant CALR functions: gains and losses

In this issue of Blood, Shide et al separate the roles of loss of a normal CALR allele and gain of a mutant CALR allele in CALR-driven essential thrombocythemia (ET).¹³ 

Approximately 1 in 4 patients with ET, a blood cancer characterized by overproduction of platelets, has a frameshift mutation in the gene encoding calreticulin (CALR).^2,3  Research to date has demonstrated a gain-of-function role for the frameshifted CALR protein in binding to the thrombopoietin receptor (TpoR), thus acting as a rogue ligand and triggering proliferation and megakaryopoiesis. However, there has been less investigation of the effects of the loss of a normal CALR allele: Calr^−/− mice are embryonic lethal because of a malformation of the heart, which prevents studying hematopoiesis in these mice.⁴ 

Shide et al generated a mouse with hematopoietic-specific deletion of the Calr allele to allow separation of gain- and loss-of-function effects of the Calr mutation. Loss of 1 or both Calr alleles in hematopoietic cells had little effect on bone marrow hematopoiesis and did not cause ET, although loss of both alleles increased extramedullary hematopoiesis. In transplantation experiments, loss of 1 or both alleles increased bone marrow repopulation in primary recipients, but only loss of a single allele maintained this advantage in secondary transplants, indicating that Calr haploinsufficiency yields a competitive advantage. The authors further demonstrate that a CALRdel52 transgene can drive a myeloproliferative neoplasm (MPN) phenotype (thrombocytosis) in transplants only when wild-type (WT) Calr is haploinsufficient. Finally, they used transcriptomic data to investigate potential mechanisms that underlie the advantage of Calr^+/− cells. Compared with WT stem/progenitor cells, they find that Calr^+/− cells have increased expression of E2F targets, suggestive of higher cell cycling. When comparing WT and Calr^+/− cells that also express transgenic CALRdel52, the authors again show upregulation of E2F target genes, in addition to an increase in stem cell self-renewal pathways and decreases in pathways responsive to the pro-inflammatory cytokines tumor necrosis factor α and interferon γ.

These results advance the field in several ways: one of these is by demonstrating a previously unknown role for normal CALR in hematopoiesis. In addition, they shed light on important questions about differences between mouse models and human patients: human patients typically show clonal disease, but mouse models to date have shown either no^5,6  or a slow-rising⁷  competitive advantage. The results of Shide et al suggest that an approach wherein 2 WT Calr alleles are maintained⁶  may show no competitive advantage, because this requires loss of a normal Calr allele. Therefore, knockin approaches may be more informative, because they have 1 WT and 1 mutant allele, thus more closely resembling patients. One mouse model with knockin of a humanized Calrdel52 allele gave rise to a strong ET phenotype but no advantage within transplants.⁵  A separate CRISPR-Cas9–based mouse model of the Calrdel52 mutation within the mouse Calr gene exhibited a lower platelet phenotype and a slow-rising competitive advantage in primary transplants, consistent with the requirement for Calr haploinsufficiency to yield a stem cell advantage.⁷  The differences between the models remain unresolved but may reflect differences in how murine and human frameshifted CALR protein binds to TpoR⁷  or other differences in the generation of the models. Furthermore, the results of Shide et al emphasize that the level of Calrdel52 expression is crucial: progression to myelofibrosis in mouse models is seen only when CALRdel52 is highly expressed via retroviral expression⁶  or homozygosity of the knocked-in Calrdel52 allele.⁵  Together, these results underscore the importance of ensuring that model systems resemble the situation in human patients, who have 1 mutant and 1 WT allele. Similarly, although results from transplant experiments can give insight into stem cell function, it is important to remember that these are artificial settings that do not resemble steady-state hematopoiesis in patients.

Going forward, more research will be necessary to understand why Calr haploinsufficiency is so critical for bestowing a competitive advantage on stem cells: although there are indications of increased cell cycling and self-renewal, no direct mechanism has been proposed. It remains to be seen whether this mechanism will rely on one of the canonical functions of CALR, such as protein chaperoning or calcium signaling,⁸  or whether the mechanism will be as surprising as the discovery of TpoR activation by frameshifted CALR protein. Because the transcriptomic data were obtained from stem/progenitor cells from transplant recipients, they may be confounded by the stresses of the transplantation protocol on both donor cells and the recipient bone marrow niche. Overall, this study is an important advance in our understanding of mutant CALR-driven ET: in addition to the well-studied role of mutant CALR as a rogue ligand for TpoR, Shide et al show a role for the loss of a WT Calr allele that warrants further investigation.