Skip to main content
Device Global Header

Subscribe
or Renew

Advanced Search

Original Article

Somatic Mutations of Calreticulin in Myeloproliferative Neoplasms

List of authors.
  • Thorsten Klampfl, Ph.D.,
  • Heinz Gisslinger, M.D.,
  • Ashot S. Harutyunyan, M.D., Ph.D.,
  • Harini Nivarthi, Ph.D.,
  • Elisa Rumi, M.D.,
  • Jelena D. Milosevic, M.Sc.,
  • Nicole C.C. Them, M.Sc.,
  • Tiina Berg, B.Sc.,
  • Bettina Gisslinger, M.Sc.,
  • Daniela Pietra, Ph.D.,
  • Doris Chen, Ph.D.,
  • Gregory I. Vladimer, Ph.D.,
  • Klaudia Bagienski, M.Sc.,
  • Chiara Milanesi, M.Sc.,
  • Ilaria Carola Casetti, M.D.,
  • Emanuela Sant'Antonio, M.D.,
  • Virginia Ferretti, Ph.D.,
  • Chiara Elena, M.D.,
  • Fiorella Schischlik, M.Sc.,
  • Ciara Cleary, M.Sc.,
  • Melanie Six, B.Sc.,
  • Martin Schalling, M.Sc.,
  • Andreas Schönegger, M.Sc.,
  • Christoph Bock, Ph.D.,
  • Luca Malcovati, M.D.,
  • Cristiana Pascutto, Ph.D.,
  • Giulio Superti-Furga, Ph.D.,
  • Mario Cazzola, M.D.,
  • and Robert Kralovics, Ph.D.

Abstract

Background

Approximately 50 to 60% of patients with essential thrombocythemia or primary myelofibrosis carry a mutation in the Janus kinase 2 gene (JAK2), and an additional 5 to 10% have activating mutations in the thrombopoietin receptor gene (MPL). So far, no specific molecular marker has been identified in the remaining 30 to 45% of patients.

Methods

We performed whole-exome sequencing to identify somatically acquired mutations in six patients who had primary myelofibrosis without mutations in JAK2 or MPL. Resequencing of CALR, encoding calreticulin, was then performed in cohorts of patients with myeloid neoplasms.

Results

Somatic insertions or deletions in exon 9 of CALR were detected in all patients who underwent whole-exome sequencing. Resequencing in 1107 samples from patients with myeloproliferative neoplasms showed that CALR mutations were absent in polycythemia vera. In essential thrombocythemia and primary myelofibrosis, CALR mutations and JAK2 and MPL mutations were mutually exclusive. Among patients with essential thrombocythemia or primary myelofibrosis with nonmutated JAK2 or MPL, CALR mutations were detected in 67% of those with essential thrombocythemia and 88% of those with primary myelofibrosis. A total of 36 types of insertions or deletions were identified that all cause a frameshift to the same alternative reading frame and generate a novel C-terminal peptide in the mutant calreticulin. Overexpression of the most frequent CALR deletion caused cytokine-independent growth in vitro owing to the activation of signal transducer and activator of transcription 5 (STAT5) by means of an unknown mechanism. Patients with mutated CALR had a lower risk of thrombosis and longer overall survival than patients with mutated JAK2.

Conclusions

Most patients with essential thrombocythemia or primary myelofibrosis that was not associated with a JAK2 or MPL alteration carried a somatic mutation in CALR. The clinical course in these patients was more indolent than that in patients with the JAK2 V617F mutation. (Funded by the MPN Research Foundation and Associazione Italiana per la Ricerca sul Cancro.)

Introduction

Philadelphia chromosome–negative myeloproliferative neoplasms include polycythemia vera, essential thrombocythemia, and primary myelofibrosis.1 A unique gain-of-function mutation in the Janus kinase 2 gene (JAK2) is found in about three quarters of patients in whom these disease entities have been diagnosed.2,3 The valine-to-phenylalanine (V617F) alteration constitutively activates JAK2, resulting in increased phosphorylation of its substrates and leading to increased cytokine responsiveness of myeloid cells. The JAK2 V617F mutation is present in approximately 95% of patients with polycythemia vera and in 50 to 60% of those with essential thrombocythemia or primary myelofibrosis.4

In addition, somatic mutations of JAK2 exon 12 are found in polycythemia vera,5,6 and activating mutations of the thrombopoietin receptor gene MPL are present in 5 to 10% of patients with essential thrombocythemia or primary myelofibrosis with nonmutated JAK2. 7,8 Thus, whereas somatic JAK2 mutations are found in nearly all patients with polycythemia vera,9 approximately one third of patients with essential thrombocythemia or primary myelofibrosis do not carry any mutation in JAK2 or MPL.7,8 Somatic mutations in other genes, such as TET2, 10,11 CBL, 12,13 EZH2, 14,15 DNMT3A, 16 and ASXL1, 17 are present in a proportion of cases of myeloproliferative neoplasms, but these can co-occur with JAK2 and MPL mutations and are found in all types of myeloid cancers.

The natural history of Philadelphia chromosome–negative myeloproliferative neoplasms is characterized not only by the occurrence of thromboembolic complications but also by a tendency toward progression to more aggressive disease, including post–polycythemia vera myelofibrosis or post–essential thrombocythemia myelofibrosis and acute myeloid leukemia or blast-phase disease.1 Progression of the disease is typically associated with the acquisition of somatic mutations in driver genes responsible for subclonal evolution.18-22 However, these mutated genes do not have a primary role in the pathogenesis of essential thrombocythemia or primary myelofibrosis associated with nonmutated JAK2 and MPL. We performed whole-exome sequencing to identify somatic mutations responsible for the initiation of disease in these subsets of myeloproliferative neoplasms.

Methods

Patients and Samples

We studied patients with Philadelphia chromosome–negative myeloproliferative neoplasms who were followed at the Medical University of Vienna in Austria, and Fondazione Istituto di Ricovero e Cura a Carattere Scientifico Policlinico San Matteo in Pavia, Italy. The study was approved by the ethics committee at each institution, and all the patients provided written informed consent.

Genomic DNA was obtained from bone marrow mononuclear cells, peripheral-blood granulocytes, or whole-blood samples. Germline control DNA was obtained from immunomagnetically purified circulating T cells.

Whole-Exome Sequencing

Whole-exome sequencing was performed in six patients with primary myelofibrosis who did not have mutations in JAK2 or MPL. Genomic DNA libraries were generated from peripheral-blood granulocyte DNA (tumor samples) and matched CD3+ T-lymphocyte DNA (control samples) with the use of the NEBNext DNA Sample Prep Reagent set (New England BioLabs). Whole-exome enrichment was performed with the use of the SureSelect Human All Exon kit (Agilent Technologies) according to the manufacturer's instructions. The libraries were sequenced with the use of a HiSeq 2000 sequencing system (Illumina). The data analysis is described in the Supplementary Appendix, available with the full text of this article at NEJM.org.

Targeted Resequencing and Mutation Screening

Genomic regions of interest were amplified by means of polymerase chain reaction (PCR). PCR products were purified, and a sequencing reaction was set up with the use of the BigDye Terminator, version 3.1, Cycle Sequencing Kit (Life Technologies). Sequencing products were analyzed with the use of a 3130xl Genetic Analyzer (Applied Biosystems). To screen for insertion and deletion mutations in CALR exon 9, we designed PCR primers spanning exon 9 and labeled the forward primer with 6-carboxyfluorescein. After PCR, the products were sized on a 3130xl Genetic Analyzer. Details of the methods and analysis are provided in the Supplementary Appendix.

Results

Whole-Exome Sequencing in Primary Myelofibrosis

Using whole-exome sequencing, we analyzed genomic DNA from granulocytes (tumor samples) and CD3+ T cells (control samples) obtained from six patients with primary myelofibrosis. Independent validation of the detected variants with the use of classic Sanger sequencing confirmed the presence of 2 to 12 somatic mutations per patient (Table S3 in the Supplementary Appendix). The only recurrently affected gene was CALR, encoding calreticulin.

Two patients had somatic deletions in exon 9 of CALR. PCR-product subcloning and sequencing revealed that Patient H_0191 had a 52-bp deletion and Patient H_0296 had a 1-bp deletion (Fig. S1 in the Supplementary Appendix). Because the 52-bp deletion in Patient H_0191 was incorrectly annotated as a 1-bp deletion coupled with a single nucleotide variant by our variant-calling analysis pipeline (Table S3 in the Supplementary Appendix), we manually reviewed the sequence alignment for this patient. We observed a misalignment of the sequence reads covering the site of mutation, owing to a repetitive element in the affected genomic region (Fig. S2 in the Supplementary Appendix).

Following up on this finding, we investigated the alignments for the remaining four patients and detected a recurrent 5-bp insertion in all four (Fig. S3 in the Supplementary Appendix). The mutations in CALR that were found by means of whole-exome sequencing were confirmed and shown to be somatic by means of Sanger sequencing of the granulocyte and matched T-lymphocyte DNA samples from all six patients.

CALR Exon 9 Mutations in Myeloproliferative Neoplasms

To estimate the prevalence of CALR mutations in myeloproliferative neoplasms, we screened a cohort of 896 patients for insertion or deletion mutations in CALR exon 9 (Fig. S4 in the Supplementary Appendix), using high-resolution sizing of fluorescent dye–labeled PCR products (Fig. S5 in the Supplementary Appendix). This cohort included 382 patients with polycythemia vera, 311 with essential thrombocythemia, and 203 with primary myelofibrosis (Table S4 in the Supplementary Appendix).

We identified 150 samples (17% of the patients) with insertions or deletions in CALR that were independently validated by means of Sanger sequencing. We did not observe CALR mutations in patients with polycythemia vera. A total of 78 patients with essential thrombocythemia (25%) and 72 with primary myelofibrosis (35%) had mutations in CALR.

Figure 1. Figure 1. Frequency of CALR Mutations in Myeloid Neoplasms.

Panel A shows the distribution of JAK2, MPL, and CALR mutations in the three classical entities of myeloproliferative neoplasms. Panel B shows the frequency of CALR mutations in various myeloid cancers. The total numbers of patients included in the analysis for a specific disease entity are shown. Panel C shows the distribution of mutations in JAK2, MPL, CALR, and SF3B1 among 24 patients with refractory anemia with ring sideroblasts associated with marked thrombocytosis (RARS-T). AML denotes acute myeloid leukemia, CML chronic myeloid leukemia, CMML chronic myelomonocytic leukemia, and MDS myelodysplastic syndrome.

All the patients were genotyped for the JAK2 V617F mutation. Patients with polycythemia vera who were negative for the JAK2 V617F mutation were tested for mutations in JAK2 exon 12. Patients with essential thrombocythemia or primary myelofibrosis who had nonmutated JAK2 were tested for mutations in MPL exon 10. The distribution of the JAK2, MPL, and CALR mutations in the three disease entities is shown in Figure 1A. All patients with mutated CALR had nonmutated JAK2 and MPL. A total of 67 patients had nonmutated JAK2 and MPL as well as nonmutated CALR exon 9. Of these 67 patients with triple-negative findings, 19 underwent Sanger sequencing for mutations in all nine exons of CALR, but no mutations were detected (data not shown).

Because CALR mutations were strongly associated with essential thrombocythemia or primary myelofibrosis with nonmutated JAK2 and MPL, we analyzed samples from an additional 211 patients in these disease categories. In the combined cohort of 1107 patients with analyzed samples, we studied 289 patients with essential thrombocythemia and nonmutated JAK2 and MPL, of whom 195 had mutated CALR (67%). Of the combined 120 patients with primary myelofibrosis and nonmutated JAK2 and MPL, 105 had a mutation in CALR (88%). In 150 patients with mutated CALR for whom we had matched T-lymphocyte DNA available, the mutations were somatic.

CALR Exon 9 Mutations in Other Myeloid Neoplasms

To investigate whether CALR mutations are present in other myeloid neoplasms, we screened 254 patients with acute myeloid leukemia, 45 with chronic myeloid leukemia, 73 with the myelodysplastic syndrome, 64 with chronic myelomonocytic leukemia, and 24 with refractory anemia with ring sideroblasts associated with marked thrombocytosis (RARS-T) for mutations in CALR exon 9. Although most of these patients had nonmutated CALR exon 9, a total of 3 patients with RARS-T had mutations in CALR (Figure 1B), all with nonmutated JAK2 and MPL (Figure 1C). Mutations in the gene encoding splicing factor 3B, subunit 1 (SF3B1) co-occurred with mutations in all three genes. Of 524 healthy study participants, 1 had a 3-bp in-frame deletion in CALR.

CALR Mutations and a Novel C-Terminal Peptide in Mutant CALR

Figure 2. Figure 2. Mutations in CALR and the Generation of a Novel C-Terminal Peptide in the Calreticulin Protein.

Panel A shows the genomic positions of all 36 mutation types detected in the current study. Red bars indicate deletions, blue letters inserted nucleotide sequences, and orange letters somatically acquired substitutions. Black lines connecting deletions or insertions with substitutions indicate their occurrence on the same allele. The red bars and black lines are drawn to scale. The most frequent mutations, type 1 and type 2, are marked. 3′UTR denotes 3′ untranslated region. Panel B is a schematic representation of the three alternative peptide sequences derived from the three reading frames across exon 9 of CALR. Each bar represents an amino acid: blue bars indicate negatively charged amino acids, red bars positively charged amino acids, and black bars stop codons. Panel C shows the specific peptide makeup of nonmutant CALR and of the two most frequently detected types of mutations. Panel D shows the relative frequencies of all 36 mutation types observed in CALR. Percentages do not sum to 100 because of rounding.

We detected a total of 36 types of somatic mutations in CALR (insertions and deletions) that caused a frameshift (Figure 2A, and Table S5 in the Supplementary Appendix). Furthermore, two patients who were positive for the JAK2 V617F mutation had in-frame germline mutations (type G1 and G2) (Table S5 in the Supplementary Appendix). All 36 types of somatic insertions or deletions resulted in a frameshift to the alternative reading frame 1 (Figure 2B). Mutations of type 1 (52-bp deletion; c.1092_1143del) and mutations of type 2 (5-bp insertion; c.1154_1155insTTGTC) accounted for 53.0% and 31.7% of all the cases with mutated CALR, respectively (Figure 2C and 2D). The other mutation types were observed at much lower frequencies, and many were detected only in a single patient (Figure 2D, and Table S5 in the Supplementary Appendix).

Because all 36 mutation types cause a frameshift to alternative reading frame 1, the resulting mutant CALR proteins share a novel amino acid sequence at the C-terminal (Table S6 in the Supplementary Appendix). The C-terminal peptide derived from alternative reading frame 1 contains a number of positively charged amino acids, whereas the nonmutant CALR C-terminal is largely negatively charged (Figure 2B).

In addition, the nonmutant calreticulin contains the endoplasmic reticulum–retention motif at the C-terminal end (lysine, aspartic acid, glutamic acid, and leucine [KDEL] amino acid sequence). The C-terminal KDEL motif is lost in all mutant variants (Table S6 in the Supplementary Appendix). Depending on the type of mutation, the mutant proteins retain varying amounts of the negatively charged amino acids of nonmutant calreticulin. The 52-bp deletions (type 1) eliminate almost all negatively charged amino acids, whereas the 5-bp insertions (type 2) retain approximately half the negatively charged amino acids (Figure 2C).

Figure 3. Figure 3. Association of CALR Mutations with Uniparental Disomies and Clonal Hierarchies in Patients with Multiple Somatic Mutations.

Panel A shows results from the polymerase-chain-reaction product-sizing analysis of CALR exon 9 in three patients with a 5-bp insertion. In the three patients, the mutated allele peak is higher than the nonmutated allele peak, indicating the presence of cells homozygous for the mutation. The plots on the right show data from Genome-Wide human single-nucleotide polymorphism (SNP) arrays, version 6.0 (Affymetrix). Each dot represents a single SNP. The x axis shows the genomic position, and the y axis the allelic status of the SNP (an allelic difference of 0 indicates heterozygous status) according to fluorescence intensity. The array data show that the three samples have clones with uniparental disomy of chromosome 19p of different sizes, which corresponds to the differences in peak heights observed in the fragment analysis of CALR. Taken together, the results indicate that homozygosity of mutated CALR is derived from duplication of a heterozygous mutation as a result of acquisition of a uniparental disomy in the three patients. Blue boxes indicate the genomic region of the uniparental disomies. Panel B shows the clonal hierarchies derived from the analysis of hematopoietic progenitor colonies. Patient H_0191 had somatic mutations in four genes. As shown in the bar chart, 51% of the colonies had mutations in CALR, GAB2, and METTL11B. The other colonies (48%) had mutations in all four genes, indicating that the mutation in the gene encoding plant homeo domain finger protein 16 (PHF16) was acquired later and gave rise to a subclone. In addition, one colony (1%) had a 1-bp deletion in CALR, in contrast to the 52-bp deletion observed in the granulocyte sample and in all other colonies from this patient. Because this colony had none of the other mutations observed in the patient, it represents an independent clone, although this conclusion is based only on a single colony. Patient H_0296 had somatic mutations in CALR, PIK3R, and C10orf71. All colonies analyzed from this patient had all three mutations. One colony showed an 18-bp deletion in CALR in addition to the 1-bp deletion observed in this patient. The two mutations were on the same allele. Mutated CALR was found in an early clone in both patients; the roles of the other mutated genes are unclear. The plots on the right side schematically depict the clonal history that gave rise to the clonal composition observed at the time of the sample. Different colors indicate the different clones, as shown in the bar charts. CFU denotes colony-forming units, and del deletion.

Given these differences, we hypothesized that type 1 and type 2 mutations may be associated with qualitatively different phenotypes. Accordingly, we found that the type 1 deletions were significantly more frequent in primary myelofibrosis than in essential thrombocythemia (P<0.001). In addition, we found that only three patients were homozygous for CALR mutations associated with uniparental disomy of chromosome 19p, and all three had a 5-bp insertion of type 2 (Figure 3A).

Early Acquisition of CALR Mutations and Stable Mutant Clones

To investigate whether mutations in CALR are acquired early or late in the clonal history of a patient, we analyzed hematopoietic-progenitor colonies from 2 patients for whom we had mutational profiles based on whole-exome sequencing. The clonal hierarchies for Patients H_0191 and H_0296 are shown in Figure 3B. We concluded that for these 2 patients, the mutations in CALR were acquired early in the major clones. We had follow-up samples available for 24 patients with mutated CALR; all these follow-up samples tested positive for the mutation as well.

Clinical Significance of CALR Mutations

Overall, we studied 1215 patients with essential thrombocythemia or primary myelofibrosis (Fig. S4 and Table S7 in the Supplementary Appendix). Of these patients, 63.4% carried the JAK2 V617F mutation, 4.4% carried activating mutations of MPL exon 10, 23.5% carried mutations of CALR exon 9, and only 8.8% had none of these clonal markers. Most of the patients with no clonal markers were clustered in the subgroup of patients with essential thrombocythemia.

We used the Wilcoxon rank-sum test to compare hematologic values in patients carrying different mutated genes. Among patients with essential thrombocythemia, those with a CALR mutation had a lower hemoglobin level, lower white-cell count, and higher platelet count at diagnosis than patients with mutated JAK2 (P<0.001 for all comparisons). Among patients with primary myelofibrosis, those with a CALR mutation had a lower white-cell count (P=0.03) and a higher platelet count (P<0.001) than patients with mutated JAK2.

Overall survival and the risk of thrombosis were analyzed only among patients carrying a mutation in JAK2, MPL, or CALR (i.e., patients with a clonal marker). Assuming that mutation status did not change with time, we performed survival analyses using the date of the initial diagnosis as the start of follow-up. Because 6 patients were excluded owing to inadequate follow-up, a total of 1102 patients were examined. The median follow-up for the entire cohort of patients with any of the three mutated genes was 5.7 years (range, 0 to 31).

Figure 4. Figure 4. Estimated Overall Survival and Cumulative Incidence of Thrombosis among Patients with CALR, JAK2, or MPL Mutations.

Outcome estimates in patients with essential thrombocythemia or primary myelofibrosis were stratified according to somatic mutation. Panel A shows the Kaplan–Meier analysis of overall survival among patients with primary myelofibrosis. Subgroups were compared with the use of the log-rank test. Patients with myelofibrosis and a somatic mutation of CALR had longer overall survival than those with the JAK2 V617F mutation (median, 21.4 years [95% confidence interval {CI}, 17.1 to 22.9] vs. 11.0 years [95% CI, 7.8 to 14.4]; P<0.001) or the MPL mutation (median, 8.2 years [95% CI, 2.0 to not reached]; P<0.001]. No significant difference in survival was observed between the JAK2 V617F and MPL subgroups. Panel B shows the Kaplan–Meier analysis of overall survival among patients with essential thrombocythemia. Subgroups were compared with the use of the log-rank test. The median value for overall survival was not reached in any subgroup. Patients with a CALR mutation had longer overall survival than those with the JAK2 V617F mutation. Overall survival at 10 years was 96.9% (95% CI, 91.7 to 98.8) in the former group, as compared with 91.1% (95% CI, 87.1 to 93.9) in the latter group (P=0.04). Panel C shows the cumulative incidence of thrombosis among patients with essential thrombocythemia. Death in the absence of the event of interest was considered a competing event, and subgroups were compared with the use of the Pepe–Mori test.23 Patients with a CALR exon 9 mutation had a lower cumulative incidence of thrombosis than those with the JAK2 V617F mutation: the cumulative incidence at 10 years was 11.0% (95% CI, 6.3 to 17.1) in the former group versus 21.0% (95% CI, 16.6 to 25.7) in the latter group (P=0.003).

As shown in Figure 4A, there was a significant difference in overall survival among the three subgroups of patients with primary myelofibrosis (P<0.001). Patients with a somatic mutation of CALR had longer overall survival than those with a JAK2 or MPL mutation (P<0.001 for both comparisons), whereas no significant difference was observed between the latter two subgroups. Among patients with essential thrombocythemia, who had much longer overall survival than those with primary myelofibrosis, there was a significant difference only between patients with a CALR mutation and those with a JAK2 mutation (P=0.04) (Figure 4B).

In a multivariate Cox regression analysis of overall survival that included type of myeloid neoplasm (essential thrombocythemia vs. primary myelofibrosis), type of mutated gene, and patient cohort (Pavia vs. Vienna) as covariates, the first two variables were found to be independent prognostic factors. As expected, primary myelofibrosis was associated with shorter overall survival, as compared with essential thrombocythemia (hazard ratio for death, 7.1; 95% confidence interval [CI], 4.9 to 10.2; P<0.001). In addition, the type of mutated gene had an independent effect on survival. As compared with patients with a CALR mutation, patients with a JAK2 mutation had a higher risk of death (hazard ratio, 3.1; 95% CI, 2.0 to 4.7; P<0.001), as did those with an MPL mutation (hazard ratio, 3.5; 95% CI, 1.8 to 6.7; P<0.001).

The cumulative incidence of thrombosis in essential thrombocythemia was calculated with a competing-risk approach,24 with death from any cause as a competing event; the curves are shown in Figure 4C, and the data are provided in Table S8 in the Supplementary Appendix. Among patients with essential thrombocythemia, those with a CALR mutation had a lower risk of thrombosis than did those with a JAK2 mutation (P=0.003); no significant differences were found between the other subgroups. It should be noted that the subgroup of patients with an MPL mutation was small.

Functional Analysis of the Type 1 CALR Mutation

Figure 5. Figure 5. Functional Analysis of CALR Type 1 Mutation.

Panel A shows the viability of cells from a murine interleukin-3–dependent pro-B cell line (Ba/F3 cells) expressing an empty vector (green fluorescent protein [GFP]), nonmutant CALR with GFP, or mutant CALR with GFP. Viability was assessed after 72 hours in the presence of an increasing concentration of interleukin-3. I bars represent standard errors. RLU denotes relative luminescence units. Panel B shows the proliferation of Ba/F3 cells expressing an empty vector, nonmutant CALR, or mutant CALR. Proliferation was assessed after 7 days in the absence of interleukin-3. Panel C shows the activation of signal transducer and activator of transcription 5 (STAT5) in response to interleukin-3. Ba/F3 cells expressing the empty vector, the nonmutant CALR, or the mutant CALR were cultured for 4 hours in serum-free medium without interleukin-3 and were then stimulated for 20 minutes with interleukin-3 at concentrations of 0.1 ng per milliliter and 1 ng per milliliter, as indicated. Western blot analysis was performed on the cell lysates with antibodies against phosphorylated STAT 5 (pSTAT5), STAT5, and CALR. An antibody against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. Panel D shows immunofluorescence staining of CALR (red) and calnexin, an endoplasmic reticulum–specific marker (green) in human embryonic kidney T cells transfected with the respective constructs. The nonmutant CALR colocalizes with calnexin in the endoplasmic reticulum (merge in yellow). However, the mutant CALR is not constrained within the endoplasmic reticulum. The nucleus is stained blue (4′,6-diamidino-2-phenylidole [DAPI]).

To study the functional effects of mutated CALR, we cloned the complementary DNA (cDNA) of the nonmutated CALR and the type 1 mutation (52-bp deletion) into a retroviral expression vector based on the murine stem-cell virus, with an internal ribosome entry site and green fluorescent protein (GFP). After retroviral production and transfection of the CALR cDNAs into the interleukin-3–dependent murine cell line Ba/F3, we sorted the transgene-positive cells by means of flow cytometry for GFP (Fig. S6 in the Supplementary Appendix). We then measured the interleukin-3–dependent proliferation of cells. Cells expressing the type 1 CALR mutation showed growth that was independent of interleukin-3 and also showed hypersensitivity to interleukin-3 (Figure 5A). When we measured the proliferation of cells in the absence of interleukin-3, only the CALR type 1 mutation showed a significant accumulation of cells (Figure 5B).

To investigate whether the interleukin-3 independence in the CALR type 1–mutated cells was caused by the activation of JAK–signal transducer and activator of transcription (JAK-STAT) signaling, we determined the sensitivity of cells to the JAK2 kinase inhibitor SAR302503. Cells expressing the nonmutated CALR or the type 1 mutation of CALR showed similar sensitivity to SAR302503, suggesting that the interleukin-3–independent growth of the mutated CALR cells depends on JAK2 or a JAK family kinase targeted by SAR302503 (Fig. S7 in the Supplementary Appendix).

To confirm this hypothesis, we examined the phosphorylation of STAT5 in the presence and absence of interleukin-3 in the control and CALR-transfected cell lines. We detected increased phosphorylation of STAT5 in the absence of interleukin-3 in the type 1 mutation of CALR and at an interleukin-3 concentration of 0.1 ng per milliliter (Figure 5C). Thus, increased activation of JAK-STAT signaling is probably responsible for the cytokine-independent growth of cells expressing the type 1 mutation of CALR.

Immunofluorescence microscopy was used to determine the cellular localization of nonmutated and type 1–mutant CALR. On overexpression in human embryonic kidney (HEK) cells, the nonmutant CALR colocalized with the endoplasmic reticulum (stained with calnexin). In the case of the type 1–mutant CALR, however, this colocalization was less prominent, probably because of the absence of the C-terminal KDEL sequence of the mutant CALR (Figure 5D).

Discussion

We identified somatic mutations in CALR in patients with primary myelofibrosis or essential thrombocythemia. CALR mutations are mutually exclusive with mutations in both JAK2 and MPL. No CALR mutations were found in patients with polycythemia vera, a myeloproliferative neoplasm that is specifically associated with JAK2 mutations. CALR mutations are the second most frequent mutation after JAK2 in myeloproliferative neoplasms. We also studied patients with other myeloid neoplasms and found CALR mutations only in a few patients with RARS-T, a myeloid neoplasm with both myelodysplastic and myeloproliferative features.25 These observations strongly support a causal relationship between CALR mutations and excessive platelet production.

Because CALR mutations are found in approximately 73% of patients who do not have mutations of JAK2 and MPL, we believe they fill in the current molecular diagnostic gap in myeloproliferative neoplasms. Altogether, less than 10% of our patients with essential thrombocythemia or primary myelofibrosis did not have a somatic mutation of JAK2, MPL, or CALR. In some of these patients, the mutated clone might have been too small to be detected with the current approaches. Rare mutated driver genes may play a role in some patients, whereas other patients might not have a clonal disease at all. This is particularly true for patients with a clinical diagnosis of essential thrombocythemia, because the differential diagnosis between clonal and reactive thrombocytosis may be difficult without a clonal marker.26 Assessment for CALR mutations markedly improves the current diagnostic approach for essential thrombocythemia and primary myelofibrosis and may be a useful addition to the World Health Organization criteria for these disorders.1

All the mutations of CALR we identified are insertion or deletion mutations in the last exon encoding the C-terminal amino acids of the protein. Most of these mutations are present in a heterozygous state and cause a frameshift to a specific alternative reading frame. This frameshift results in the replacement of the C-terminal negatively charged amino acids of calreticulin by a positively charged polypeptide rich in arginine and methionine. The last four amino acids of calreticulin (KDEL) contain the endoplasmic reticulum–retention signal. This signal is absent in mutant calreticulin. Consequently, mutant calreticulin has an altered subcellular localization. Because the negatively charged C-terminal domain of calreticulin is the low-affinity, high-capacity, Ca2+-binding domain, the Ca2+-binding function of the mutant protein may be impaired. The presence of the peptide sequence derived from the alternative reading frame at the C-terminal domain of mutated CALR offers an opportunity for immunologic targeting because it represents a cancer-specific epitope.

To analyze the oncogenic capability of the mutant calreticulin, we generated Ba/F3 cells with overexpression of the nonmutant and the type 1 mutant CALR (52-bp deletion). The Ba/F3 cells expressing the 52-bp deletion showed cytokine-independent proliferation. However, the growth of Ba/F3 cells expressing nonmutant and mutant calreticulin was suppressed equally on treatment with a JAK2 kinase inhibitor, suggesting that the JAK-STAT pathway is required for cytokine-independent proliferation induced by mutant calreticulin.

In accordance with this finding, we detected increased phosphorylation of STAT5 in the Ba/F3 cells with the 52-bp deletion, both in the absence and in the presence of stimulation by interleukin-3. Calreticulin, Ca2+, and calmodulin have previously been shown to modulate the activity of Stats. The calreticulin complex with endoplasmic reticulum protein 57, in the endoplasmic reticulum, suppresses the phosphorylation and transcriptional activity of Stat3 in mouse embryonic fibroblasts.27 Moreover, inhibition of the calcium-calmodulin–dependent protein kinase II γ results in reduced levels of phosphorylated Stat1, Stat3, and Stat5.28 Overexpression of calreticulin attenuates interferon α–induced Stat1 phosphorylation, resulting in interferon resistance.29

Studies are required to elucidate the mechanism for activation of the JAK-STAT pathway by the mutant calreticulin in myeloid cells. The involvement of the JAK-STAT signaling pathway in patients who are positive for a CALR mutation may also explain the effectiveness of JAK2-inhibitor therapy for primary myelofibrosis. However, our results indicate that the JAK2 inhibitors may not be selective for cells expressing the mutated CALR, as compared with the CALR nonmutated cells.

Although our analyses of clinical outcome are retrospective, they strongly suggest that CALR-positive myeloproliferative neoplasms have a more benign clinical course than the corresponding disorders associated with JAK2 or MPL mutations. Owing to the small number of patients with an MPL mutation, the more reliable comparisons are those between patients with a JAK2 mutation and those with a CALR mutation. Our results clearly show that patients with a CALR mutation have a lower risk of thrombosis and have longer overall survival than those with a JAK2 mutation. The lower incidence of thromboembolic complications might be related to our observation that patients with a CALR mutation had lower hemoglobin levels and lower white-cell counts than those with the JAK2 mutation. Longer overall survival was observed among patients with primary myelofibrosis or essential thrombocythemia carrying a CALR mutation than among those with a JAK2 mutation; in particular, it was much more relevant in primary myelofibrosis, which confirms previous findings in patients with and those without a JAK2 mutation.8,30

From a practical point of view, the different effects of mutated genes might be incorporated into existing prognostic scoring systems for primary myelofibrosis and essential thrombocythemia31,32 and might also guide therapeutic decision making. CALR molecular characterization may become a key component of the clinical management of essential thrombocythemia and primary myelofibrosis.

Funding and Disclosures

Supported by grants from the MPN Research Foundation (to Dr. Kralovics) and Associazione Italiana per la Ricerca sul Cancro (Special Program Molecular Clinical Oncology 5 per Mille, project 1005, to Dr. Cazzola).

Drs. Kralovics, Klampfl, and Gisslinger report holding pending patent applications (EP 13.18.6939.8 and EP 13.18.4632.1) regarding the use of calreticulin gene mutations for the diagnosis of diseases and targeting for therapy for myeloproliferative neoplasms. No other potential conflict of interest relevant to this article was reported.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

This article was published on December 10, 2013, at NEJM.org.

We thank Zdenek Racil, Michael Steurer, Ingrid Pabinger, Elisabeth Koller, and Sonja Pavlovic for sample contribution.

Author Affiliations

From CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences (T.K., A.S.H., H.N., J.D.M., N.C.C.T., T.B., D.C., G.I.V., K.B., F.S., C.C., M. Six, A.S., C.B., G.S.-F., R.K.) and the Department of Internal Medicine I, Division of Hematology and Blood Coagulation, Medical University of Vienna (H.G., B.G., M. Schalling, R.K.) — both in Vienna; and the Department of Hematology Oncology, Fondazione Istituto di Ricovero e Cura a Carattere Scientifico Policlinico San Matteo (E.R., D.P., C.M., I.C.C., E.S., V.F., C.E., L.M., C.P., M.C.), and the Department of Molecular Medicine, University of Pavia (L.M., M.C.) — both in Pavia, Italy.

Address reprint requests to Dr. Kralovics at CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Lazarettgasse 14, AKH BT25.3, 1090 Vienna, Austria, or at .

Supplementary Material

References (32)

  1. 1. Swerdlow SH, Campo E, Harris NL, et al. WHO classification of tumours of haematopoietic and lymphoid tissues. Lyon, France: International Agency for Research on Cancer, 2008.

  2. 2. James C, Ugo V, Le Couedic JP, et al. A unique clonal JAK2 mutation leading to constitutive signalling causes polycythaemia vera. Nature 2005;434:1144-1148

  3. 3. Kralovics R, Passamonti F, Buser AS, et al. A gain-of-function mutation of JAK2 in myeloproliferative disorders. N Engl J Med 2005;352:1779-1790

  4. 4. Campbell PJ, Green AR. The myeloproliferative disorders. N Engl J Med 2006;355:2452-2466

  5. 5. Scott LM, Tong W, Levine RL, et al. JAK2 exon 12 mutations in polycythemia vera and idiopathic erythrocytosis. N Engl J Med 2007;356:459-468

  6. 6. Passamonti F, Elena C, Schnittger S, et al. Molecular and clinical features of the myeloproliferative neoplasm associated with JAK2 exon 12 mutations. Blood 2011;117:2813-2816

  7. 7. Pikman Y, Lee BH, Mercher T, et al. MPLW515L is a novel somatic activating mutation in myelofibrosis with myeloid metaplasia. PLoS Med 2006;3:e270-e270

  8. 8. Rumi E, Pietra D, Guglielmelli P, et al. Acquired copy-neutral loss of heterozygosity of chromosome 1p as a molecular event associated with marrow fibrosis in MPL-mutated myeloproliferative neoplasms. Blood 2013;121:4388-4395

  9. 9. Passamonti F, Rumi E, Pietra D, et al. A prospective study of 338 patients with polycythemia vera: the impact of JAK2 (V617F) allele burden and leukocytosis on fibrotic or leukemic disease transformation and vascular complications. Leukemia 2010;24:1574-1579

  10. 10. Delhommeau F, Dupont S, Della Valle V, et al. Mutation in TET2 in myeloid cancers. N Engl J Med 2009;360:2289-2301

  11. 11. Tefferi A, Pardanani A, Lim KH, et al. TET2 mutations and their clinical correlates in polycythemia vera, essential thrombocythemia and myelofibrosis. Leukemia 2009;23:905-911

  12. 12. Grand FH, Hidalgo-Curtis CE, Ernst T, et al. Frequent CBL mutations associated with 11q acquired uniparental disomy in myeloproliferative neoplasms. Blood 2009;113:6182-6192

  13. 13. Sanada M, Suzuki T, Shih LY, et al. Gain-of-function of mutated C-CBL tumour suppressor in myeloid neoplasms. Nature 2009;460:904-908

  14. 14. Ernst T, Chase AJ, Score J, et al. Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat Genet 2010;42:722-726

  15. 15. Guglielmelli P, Biamonte F, Score J, et al. EZH2 mutational status predicts poor survival in myelofibrosis. Blood 2011;118:5227-5234

  16. 16. Abdel-Wahab O, Pardanani A, Rampal R, Lasho TL, Levine RL, Tefferi A. DNMT3A mutational analysis in primary myelofibrosis, chronic myelomonocytic leukemia and advanced phases of myeloproliferative neoplasms. Leukemia 2011;25:1219-1220

  17. 17. Carbuccia N, Murati A, Trouplin V, et al. Mutations of ASXL1 gene in myeloproliferative neoplasms. Leukemia 2009;23:2183-2186

  18. 18. Kralovics R. Genetic complexity of myeloproliferative neoplasms. Leukemia 2008;22:1841-1848

  19. 19. Klampfl T, Harutyunyan A, Berg T, et al. Genome integrity of myeloproliferative neoplasms in chronic phase and during disease progression. Blood 2011;118:167-176

  20. 20. Harutyunyan A, Klampfl T, Cazzola M, Kralovics R. p53 Lesions in leukemic transformation. N Engl J Med 2011;364:488-490

  21. 21. Vannucchi AM, Lasho TL, Guglielmelli P, et al. Mutations and prognosis in primary myelofibrosis. Leukemia 2013;27:1861-1869

  22. 22. Lasho TL, Jimma T, Finke CM, et al. SRSF2 mutations in primary myelofibrosis: significant clustering with IDH mutations and independent association with inferior overall and leukemia-free survival. Blood 2012;120:4168-4171

  23. 23. Pepe MS, Mori M. Kaplan-Meier, marginal or conditional probability curves in summarizing competing risks failure time data? Stat Med 1993;12:737-751

  24. 24. Prentice RL, Kalbfleisch JD, Peterson AV Jr, Flournoy N, Farewell VT, Breslow NE. The analysis of failure times in the presence of competing risks. Biometrics 1978;34:541-554

  25. 25. Malcovati L, Della Porta MG, Pietra D, et al. Molecular and clinical features of refractory anemia with ringed sideroblasts associated with marked thrombocytosis. Blood 2009;114:3538-3545

  26. 26. Schafer AI. Thrombocytosis. N Engl J Med 2004;350:1211-1219

  27. 27. Coe H, Jung J, Groenendyk J, Prins D, Michalak M. ERp57 modulates STAT3 signaling from the lumen of the endoplasmic reticulum. J Biol Chem 2010;285:6725-6738

  28. 28. Si J, Collins SJ. Activated Ca2+/calmodulin-dependent protein kinase IIgamma is a critical regulator of myeloid leukemia cell proliferation. Cancer Res 2008;68:3733-3742

  29. 29. Yue X, Wang H, Zhao F, et al. Hepatitis B virus-induced calreticulin protein is involved in IFN resistance. J Immunol 2012;189:279-286

  30. 30. Campbell PJ, Griesshammer M, Dohner K, et al. V617F mutation in JAK2 is associated with poorer survival in idiopathic myelofibrosis. Blood 2006;107:2098-2100

  31. 31. Passamonti F, Cervantes F, Vannucchi AM, et al. A dynamic prognostic model to predict survival in primary myelofibrosis: a study by the IWG-MRT (International Working Group for Myeloproliferative Neoplasms Research and Treatment). Blood 2010;115:1703-1708

  32. 32. Passamonti F, Thiele J, Girodon F, et al. A prognostic model to predict survival in 867 World Health Organization-defined essential thrombocythemia at diagnosis: a study by the International Working Group on Myelofibrosis Research and Treatment. Blood 2012;120:1197-1201

Citing Articles (1271)

Only the 1000 most recent citing articles are listed here.

    Letters

    Comment (1)

    Figures/Media

    1. Figure 1. Frequency of CALR Mutations in Myeloid Neoplasms.
      Figure 1. Frequency of CALR Mutations in Myeloid Neoplasms.

      Panel A shows the distribution of JAK2, MPL, and CALR mutations in the three classical entities of myeloproliferative neoplasms. Panel B shows the frequency of CALR mutations in various myeloid cancers. The total numbers of patients included in the analysis for a specific disease entity are shown. Panel C shows the distribution of mutations in JAK2, MPL, CALR, and SF3B1 among 24 patients with refractory anemia with ring sideroblasts associated with marked thrombocytosis (RARS-T). AML denotes acute myeloid leukemia, CML chronic myeloid leukemia, CMML chronic myelomonocytic leukemia, and MDS myelodysplastic syndrome.

    2. Figure 2. Mutations in CALR and the Generation of a Novel C-Terminal Peptide in the Calreticulin Protein.
      Figure 2. Mutations in CALR and the Generation of a Novel C-Terminal Peptide in the Calreticulin Protein.

      Panel A shows the genomic positions of all 36 mutation types detected in the current study. Red bars indicate deletions, blue letters inserted nucleotide sequences, and orange letters somatically acquired substitutions. Black lines connecting deletions or insertions with substitutions indicate their occurrence on the same allele. The red bars and black lines are drawn to scale. The most frequent mutations, type 1 and type 2, are marked. 3′UTR denotes 3′ untranslated region. Panel B is a schematic representation of the three alternative peptide sequences derived from the three reading frames across exon 9 of CALR. Each bar represents an amino acid: blue bars indicate negatively charged amino acids, red bars positively charged amino acids, and black bars stop codons. Panel C shows the specific peptide makeup of nonmutant CALR and of the two most frequently detected types of mutations. Panel D shows the relative frequencies of all 36 mutation types observed in CALR. Percentages do not sum to 100 because of rounding.

    3. Figure 3. Association of CALR Mutations with Uniparental Disomies and Clonal Hierarchies in Patients with Multiple Somatic Mutations.
      Figure 3. Association of CALR Mutations with Uniparental Disomies and Clonal Hierarchies in Patients with Multiple Somatic Mutations.

      Panel A shows results from the polymerase-chain-reaction product-sizing analysis of CALR exon 9 in three patients with a 5-bp insertion. In the three patients, the mutated allele peak is higher than the nonmutated allele peak, indicating the presence of cells homozygous for the mutation. The plots on the right show data from Genome-Wide human single-nucleotide polymorphism (SNP) arrays, version 6.0 (Affymetrix). Each dot represents a single SNP. The x axis shows the genomic position, and the y axis the allelic status of the SNP (an allelic difference of 0 indicates heterozygous status) according to fluorescence intensity. The array data show that the three samples have clones with uniparental disomy of chromosome 19p of different sizes, which corresponds to the differences in peak heights observed in the fragment analysis of CALR. Taken together, the results indicate that homozygosity of mutated CALR is derived from duplication of a heterozygous mutation as a result of acquisition of a uniparental disomy in the three patients. Blue boxes indicate the genomic region of the uniparental disomies. Panel B shows the clonal hierarchies derived from the analysis of hematopoietic progenitor colonies. Patient H_0191 had somatic mutations in four genes. As shown in the bar chart, 51% of the colonies had mutations in CALR, GAB2, and METTL11B. The other colonies (48%) had mutations in all four genes, indicating that the mutation in the gene encoding plant homeo domain finger protein 16 (PHF16) was acquired later and gave rise to a subclone. In addition, one colony (1%) had a 1-bp deletion in CALR, in contrast to the 52-bp deletion observed in the granulocyte sample and in all other colonies from this patient. Because this colony had none of the other mutations observed in the patient, it represents an independent clone, although this conclusion is based only on a single colony. Patient H_0296 had somatic mutations in CALR, PIK3R, and C10orf71. All colonies analyzed from this patient had all three mutations. One colony showed an 18-bp deletion in CALR in addition to the 1-bp deletion observed in this patient. The two mutations were on the same allele. Mutated CALR was found in an early clone in both patients; the roles of the other mutated genes are unclear. The plots on the right side schematically depict the clonal history that gave rise to the clonal composition observed at the time of the sample. Different colors indicate the different clones, as shown in the bar charts. CFU denotes colony-forming units, and del deletion.

    4. Figure 4. Estimated Overall Survival and Cumulative Incidence of Thrombosis among Patients with CALR, JAK2, or MPL Mutations.
      Figure 4. Estimated Overall Survival and Cumulative Incidence of Thrombosis among Patients with CALR, JAK2, or MPL Mutations.

      Outcome estimates in patients with essential thrombocythemia or primary myelofibrosis were stratified according to somatic mutation. Panel A shows the Kaplan–Meier analysis of overall survival among patients with primary myelofibrosis. Subgroups were compared with the use of the log-rank test. Patients with myelofibrosis and a somatic mutation of CALR had longer overall survival than those with the JAK2 V617F mutation (median, 21.4 years [95% confidence interval {CI}, 17.1 to 22.9] vs. 11.0 years [95% CI, 7.8 to 14.4]; P<0.001) or the MPL mutation (median, 8.2 years [95% CI, 2.0 to not reached]; P<0.001]. No significant difference in survival was observed between the JAK2 V617F and MPL subgroups. Panel B shows the Kaplan–Meier analysis of overall survival among patients with essential thrombocythemia. Subgroups were compared with the use of the log-rank test. The median value for overall survival was not reached in any subgroup. Patients with a CALR mutation had longer overall survival than those with the JAK2 V617F mutation. Overall survival at 10 years was 96.9% (95% CI, 91.7 to 98.8) in the former group, as compared with 91.1% (95% CI, 87.1 to 93.9) in the latter group (P=0.04). Panel C shows the cumulative incidence of thrombosis among patients with essential thrombocythemia. Death in the absence of the event of interest was considered a competing event, and subgroups were compared with the use of the Pepe–Mori test.23 Patients with a CALR exon 9 mutation had a lower cumulative incidence of thrombosis than those with the JAK2 V617F mutation: the cumulative incidence at 10 years was 11.0% (95% CI, 6.3 to 17.1) in the former group versus 21.0% (95% CI, 16.6 to 25.7) in the latter group (P=0.003).

    5. Figure 5. Functional Analysis of CALR Type 1 Mutation.
      Figure 5. Functional Analysis of CALR Type 1 Mutation.

      Panel A shows the viability of cells from a murine interleukin-3–dependent pro-B cell line (Ba/F3 cells) expressing an empty vector (green fluorescent protein [GFP]), nonmutant CALR with GFP, or mutant CALR with GFP. Viability was assessed after 72 hours in the presence of an increasing concentration of interleukin-3. I bars represent standard errors. RLU denotes relative luminescence units. Panel B shows the proliferation of Ba/F3 cells expressing an empty vector, nonmutant CALR, or mutant CALR. Proliferation was assessed after 7 days in the absence of interleukin-3. Panel C shows the activation of signal transducer and activator of transcription 5 (STAT5) in response to interleukin-3. Ba/F3 cells expressing the empty vector, the nonmutant CALR, or the mutant CALR were cultured for 4 hours in serum-free medium without interleukin-3 and were then stimulated for 20 minutes with interleukin-3 at concentrations of 0.1 ng per milliliter and 1 ng per milliliter, as indicated. Western blot analysis was performed on the cell lysates with antibodies against phosphorylated STAT 5 (pSTAT5), STAT5, and CALR. An antibody against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. Panel D shows immunofluorescence staining of CALR (red) and calnexin, an endoplasmic reticulum–specific marker (green) in human embryonic kidney T cells transfected with the respective constructs. The nonmutant CALR colocalizes with calnexin in the endoplasmic reticulum (merge in yellow). However, the mutant CALR is not constrained within the endoplasmic reticulum. The nucleus is stained blue (4′,6-diamidino-2-phenylidole [DAPI]).

      More Research