The classical, BCR-ABL1 negative myeloproliferative neoplasms (MPN) are polycythemia vera (PV), essential thrombocythemia (ET) and primary myelofibrosis (PMF). The most common genetic alteration in MPN is the JAK2-V617F mutation detected in 95% of PV patients and in 50-60% of patients with ET or PMF. Mutations in exon 12 of JAK2 and in the thrombopoietin receptor gene MPL are found in an additional 5-10% of the cases. In recent years a number of other genes were shown to be affected in MPN. However, these mutations are not mutually exclusive with JAK2 and MPL mutations and are also found in other myeloid malignancies. A specific molecular marker for the remaining 40% of ET or PMF patients with wild type JAK2 and MPL is still unknown.

We used whole-exome sequencing to identify novel mutations in PMF patients with wild type JAK2 and MPL. The analysis revealed recurrent somatic insertions and deletions in CALR encoding for calreticulin. All detected mutations resulted in a frameshift and clustered in the last exon (exon 9) of the gene. Following up on this finding we developed a PCR based assay to screen 1107 MPN patients for insertion/deletion mutations in exon 9 of CALR. No mutations were detected in PV. In ET and PMF CALR mutations were mutually exclusive with mutant JAK2 and mutant MPL. Of the patients with wild type JAK2 and MPL, 67% ET and 88% PMF had mutant CALR. We also tested 19 patients with wild type CALR-exon 9 for mutations in the other exons of the gene, but all were negative. Furthermore we did not find CALR-exon 9 mutations in 254 patients with de novo acute myeloid leukemia, 45 with chronic myeloid leukemia, 73 with myelodysplastic syndrome or 64 with chronic myelomonocytic leukemia. Out of 24 patients with refractory anemia with ringed sideroblasts associated with marked thrombocytosis (RARS-T), 3 patients carried CALR mutations. These patients were wild type for JAK2 and MPL.

In total we detected 36 different types of mutations in CALR. A 52 bp deletion and a 5 bp insertion were the most prominent types found in 53% and 32% of all cases with mutant CALR. All 36 types of mutations result in a frameshift to the same alternative reading frame, generating a novel C-terminus of the mutated protein. The wild type C-terminal region of CALR contains a high-capacity calcium-binding domain and is highly negatively charged. As a consequence of the frameshift mutations the negatively charged amino acids are replaced by both neutral and positively charged amino acids. In addition, an endoplasmic reticulum retention signal present in the wild type protein is lost in the mutant variants. Expression in HEK cells showed that wild type CALR localizes in the endoplasmic reticulum, whereas this localization is less prominent in cells expressing mutant CALR. This observation is in line with the loss of the endoplasmic reticulum retention signal in the mutant protein.

Overexpression of the most common CALR mutation (a 52 bp deletion) in interleukin-3 (IL-3) dependent Ba/F3 cells led to IL-3-independent growth and hypersensitivity to IL-3. Cells overexpressing the mutant were sensitive to the JAK-family kinase inhibitor SAR302503 and showed elevated STAT5 phosphorylation in absence of IL-3. This indicates that JAK-STAT signaling is involved in the observed cytokine independent growth of mutant CALR expressing Ba/F3 cells.

ET and PMF patients with mutant CALR present with lower white blood cell counts (P<0.001 for ET, P=0.027 for PMF) and elevated platelet levels (P<0.001 in both entities) compared to patients with mutant JAK2. In both disease entities patients with mutant CALR show significantly better overall survival than patients with mutant JAK2 (P=0.043 in ET, P<0.001 in PMF). ET patients with mutant CALR had a lower risk of thrombosis in comparison to those with mutant JAK2 (P=0.003).

Mutant CALR is a novel, specific molecular marker detected in the majority of MPN patients negative for JAK2 and MPL mutations. Use of this marker in the clinic is expected to improve diagnostic and therapeutic decision-making in MPN.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution