Calreticulin (CALR) mutations were recently described in JAK2 and MPL unmutated primary myelofibrosis (PMF) and essential thrombocythemia. In the current study, we compared the clinical, cytogenetic and molecular features of patients with PMF with or without CALR, JAK2 or MPL mutations. Among 254 study patients, 147 (58%) harbored JAK2, 63 (25%) CALR and 21 (8.3%) MPL mutations; 22 (8.7%) patients were negative for all three mutations, whereas one patient expressed both JAK2 and CALR mutations. Study patients were also screened for ASXL1 (31%), EZH2 (6%), IDH (4%), SRSF2 (12%), SF3B1 (7%) and U2AF1 (16%) mutations. In univariate analysis, CALR mutations were associated with younger age (P<0.0001), higher platelet count (P<0.0001) and lower DIPSS-plus score (P=0.02). CALR-mutated patients were also less likely to be anemic, require transfusions or display leukocytosis. Spliceosome mutations were infrequent (P=0.0001) in CALR-mutated patients, but no other molecular or cytogenetic associations were evident. In multivariable analysis, CALR mutations had a favorable impact on survival that was independent of both DIPSS-plus risk and ASXL1 mutation status (P=0.001; HR 3.4 for triple-negative and 2.2 for JAK2-mutated). Triple-negative patients also displayed inferior LFS (P=0.003). The current study identifies ‘CALR–ASXL1+’ and ‘triple-negative’ as high-risk molecular signatures in PMF.
Calreticulin is a functionally complex Ca2+-binding protein localized primarily in the endoplasmic reticulum (ER) but is also found in the nucleus, cell membranes and extracellular matrix.1 Calreticulin gene (CALR) is located on chromosome 19p13.2 and contains 9 exons and spans 4.2 kb. CALR knockout mice are born dead and display impaired cardiac development, whereas post-natal overexpression also leads to cardiac defects.2, 3 Calreticulin protein consists of three domains: amino terminal N-domain (residues 1–180), central proline-rich P-domain (residues 181–290) and carboxyl terminal C-domain (residues 291–400). Functionally, calreticulin is believed to participate in Ca2+homeostasis, handling of misfolded proteins, cell adhesion, immune response to caner, phagocytosis and signaling.4, 5, 6, 7, 8, 9, 10, 11, 12, 13 Abnormal expression of calreticulin has been described in various cancers.12, 14, 15, 16 CALR promoter point mutations have previously been described in schizoaffective disorders.17
In December, 2013, two groups reported novel CARL mutations (exon 9 deletions and insertions) in JAK2 or MPL unmutated primary myelofibrosis (PMF) and essential thrombocythemia (ET).18, 19 In one of the two studies,19 among 1107 patient samples with myeloproliferative neoplasms (MPN), CALR mutations were not seen in polycythemia vera (PV; n=382) but seen in ET (25%; n=311) and PMF (35%; n=203). CALR mutations were mutually exclusive of JAK2 or MPL mutations. Accordingly, the authors included 211 additional JAK2/MPL-unmutated cases and found CALR mutational frequencies in JAK2/MPL-negative disease to be 67% in ET (n=289) and 88% in PMF (n=120). CALR mutations were not detected in acute myeloid leukemia (AML; n=254), myelodysplastic syndromes (MDS; n=73), chronic myelomonocytic leukemia (CMML; n=64) or chronic myeloid leukemia (CML; n=45). Three patients (13%) with RARS-T (n=24) displayed CALR mutations and all three were JAK2/MPL-negative. In ET, CALR, compared with JAK2, mutations were associated with lower hemoglobin level, lower leukocyte count, higher platelet count, lower risk of thrombosis and better survival. In PMF, the association was with lower leukocyte count, higher platelet count and better survival.19
In the second study,18 CALR mutations were first identified in 26 of 151 patients with MPN, including 48 PV, 62 ET and 39 PMF. As was the case in the aforementioned study, CALR mutations occurred only in JAK2/MPL-unmutated cases (26 of 31 such cases with ET or PMF). In a follow-up study, none of 217 patients with PV or 41 with post-PV MF harbored CALR mutations. These mutations were also absent in JAK2/MPL-mutated ET (n=138), post-ET MF (n=18) or PMF (n=97). On the other hand, mutational frequencies were 71%, 56% and 86% in JAK2/MPL-unmutated ET (n=112), PMF (n=32) or post-ET MF (n=14), respectively. The authors also detected CALR mutations in MDS (8.3%; n=120), CMML (3%; n=33), atypical CML (3.4%; n=29) but not in CML (n=28), AML (n=48), systemic mastocytosis (n=114), eosinophilic disorder (n=2), ‘idiopathic’ erythrocytosis (n=5), transient abnormal myelopoiesis (n=10), lymphoid malignancies (n=287) or solid tumors (n=502). In ET, CALR, compared with JAK2, mutations were associated with higher platelet counts, lower hemoglobin level and increased incidence of fibrotic transformation.
All CALR mutations seen so far in MPN involve exon 9 and were somatic insertions or deletions; two mutation variants (type 1 and type 2)19 were the most frequent: type 1 (L367fs*46) resulted from 52 bp deletion and was more frequent in PMF and type 2 (K385fs*47) resulted from 5-bp TTGTC insertion. In one of the two studies, these two common variants accounted for 53 and 32% of all mutant CALR, and only three patients were homozygous for CALR mutations and all three displayed K385fs*47.19 Several other distinct variants were seen infrequently. All CALR mutations resulted in one base pair reading frame shift and a significantly altered C-terminal that includes loss of most of the acidic domain and KDEL signal (ER retention motif comprised of lysine, aspartic acid, glutamic acid and leucine). Additional laboratory studies revealed the stem cell origination of CALR mutations and did not reveal altered distribution of the mutant protein in the ER or Golgi apparatus.18 Overexpression of the most frequent CALR deletion in Ba/F3 cell lines caused cytokine-independent growth and activation of signal transducer and activator of transcription 5 that was sensitive to pharmacologic JAK2 inhibition.19
The current study of 254 Mayo Clinic patients defines the clinical, cytogenetic and molecular correlates of CALR mutations in PMF and compares overall and leukemia-free survival in the absence (triple-negative) or presence of CALR, JAK2 or MPL mutations. The intent is to confirm the seminal observations from the above-mentioned two studies and provide additional details in terms of diagnostic and prognostic relevance.
Materials and Methods
The current study was approved by the Mayo Clinic institutional review board. All patients provided informed written consent for study sample collection as well as permission for its use in research. Inclusion to the current study required availability of archived peripheral blood or bone marrow sample collected at the time of diagnosis or first referral; a total of 254 PMF patients met these stipulations. The diagnoses of PMF and leukemic transformation were according to the World Health Organization criteria.20 Unfavorable karyotype designation and Dynamic International Prognostic Scoring System-plus (DIPSS-plus) risk categorization were as previously described.21, 22
Screening for JAK2, MPL, ASXL1, EZH2, IDH1, IDH2 and spliceosome (SRSF2, SF3B1, U2AF1) mutations was as previously described.23, 24, 25, 26, 27 For CALR mutation screening, genomic DNA was isolated from bone marrow and/or granulocyte-enriched peripheral blood according to Mayo internal guidelines. Oligonucleotide primers targeting exon 9 of CALR were used to amplify a 377 bp product: (CALR Forward 5′-CTGGCACCATCTTTGACAACTT-3′) (CALR Reverse 5′–GGCCTCTCTACAGCTCGTC-3′). Amplification final concentrations were as follows: 1 × of 10 × PCR Buffer (Roche-Applied-Science, Indianapolis, IN, USA), 0.5 U Taq Polymerase (Roche-Applied-Science), 400 nM each of forward and reverse oligo primer, 300 μM of each dNTP, 150 ng of DNA template and water to a final reaction volume of 50 μl. Cycling parameters consisted of an initial denaturation at 94 °C for 2 min; 40 cycles of denaturation at 94 °C for 15 s, annealing at 56 °C for 30 s, and extension at 72 °C for 45 s; and final extension at 72 °C for 1 min. PCR products were purified (QIAquick PCR Purification Kit, Qiagen, Valencia, CA, USA) and subjected to bidirectional sequencing. Mutations were identified using Mutation Surveyor Software (SoftGenetics, LLC, State College, PA, USA).
All statistical analyses considered clinical and laboratory parameters obtained at the time of diagnosis or first referral, which coincided, in all instances, with time of sample collection for mutation analysis. Differences in the distribution of continuous variables between categories were analyzed by either Mann–Whitney (for comparison of two groups) or Kruskal–Wallis (comparison of three or more groups) test. Patient groups with nominal variables were compared by chi-square test. Overall survival analysis was considered from the date of diagnosis or first referral (that is, date of sample collection) to date of death (uncensored) or last contact (censored). Date of leukemic transformation replaced date of death, as the uncensored variable, for estimating leukemia-free survival. Overall and leukemia-free survival curves were prepared by the Kaplan–Meier method and compared by the log-rank test. The Cox proportional hazard regression model was used for multivariable analysis. P-values less than 0.05 were considered significant. The Stat View (SAS Institute, Cary, NC, USA) statistical package was used for all calculations.
A total of 254 patients (median age 64 years; 65% males) with WHO-defined PMF was included in the current study, and all analyses were performed with clinical and laboratory parameters at the time of first Mayo Clinic referral. DIPSS-plus risk distributions were 32% high, 37% intermediate-2, 18% intermediate-1 and 13% low. Transfusion need and platelet count <100 × 109/l were recorded in 33 and 20% of patients, respectively. Cytogenetic studies showed clonal changes in 36% of patients and 10% had unfavorable karyotype. Mutational frequencies for ASXL1 (31%), EZH2 (6%), IDH (4%), SRSF2 (12%), SF3B1 (7%) and U2AF1 (16%) were as expected.
Among the 254 study patients, CALR mutations were detected in 63 (25%), JAK2 in 147 (58%) and MPL in 21 (8.3%). Twenty-two (8.7%) patients were negative for all three mutations (CALR, JAK2 and MPL), whereas one patient displayed both CALR and JAK2 mutations (Figure 1). CALR mutational frequency was 74% among those not mutated for either JAK2 or MPL (n=85). Among the 22 triple-negative patients, other clonal markers were present in 18 of them including ASXL1 mutations in 7, spliceosome mutations in 4 and abnormal karyotype in 7. The incidence of triple-negative cases was 26% among JAK2/MPL-unmutated cases. There were 49 (78%) type 1 (L367fs*46) and 9 (14%) type 2 (K385fs*47) mutations and the remaining included L367fs*51, K368fs*53, E386fs*46, K368fs*51 and K374fs*56. In contrast, in a separate cohort of 96 ET patients from the University of Varese, Italy (data not shown), CALR mutations were detected in 17 patients (18%) and included 7 with type 2 (K385fs*47) and 5 type 1 (L367fs*46), suggesting a higher frequency of type 1 mutations in PMF as opposed to ET. However, in another cohort of 51 PMF patients from the same institution, CALR mutations were seen in 17 (33%) and the frequency of K385fs*47 (n=7) and L367fs*46 (n=6) was similar (data not shown).
Clinical, cytogenetic and molecular correlates
Table 1 lists the presenting clinical characteristics of 253 study patients stratified by their mutation profile: JAK2 (n=147; 58%), CALR (n=63; 25%) and MPL (n=21; 8.3%) mutated and triple-negative; n=22; 8.7%); one patient displayed both CALR and JAK2 mutations (Figure 1) and was not included in Table 1. In univariate analysis, CALR mutations were associated with younger age (P<0.0001), higher platelet count (P<0.0001) and thrombocytosis (P=0.04), lower DIPSS-plus score (P=0.001) and were less likely to be red cell transfusion dependent (P=0.0005) or display moderate-to-severe anemia (P=0.002), thrombocytopenia (P=0.001) or leukocytosis (P=0.03). Spliceosome mutations were infrequent in CALR-mutated patients (P=0.0001), but no other molecular or cytogenetic associations were evident (Table 1).
Impact on overall and leukemia-free survival
Univariate analysis disclosed significant survival differences between CALR, JAK2, MPL-mutated and triple-negative patients (P<0.0001; Figure 2); triple-negative (hazards ratio (HR): 3.6, 95% confidence interval (CI): 1.9–6.7) and JAK2-mutated (HR 2.6, 95% CI: 1.6–4.0) patients had significantly shorter survival than those with CALR mutations and the difference was of borderline significance for MPL vs CALR-mutated cases (HR 1.7, 95% CI: 0.9–3.3). When adjusted for age, the CALR survival advantage remained significant against triple-negative (P=0.0008) and JAK2-mutated (P=0.01) but not MPL-mutated (P=0.22) cases. Similar results were seen after adjustment for each one of mutations previously identified as being prognostically relevant: ASXL1, EZH2, SRSF2 and IDH1/2. When the analysis included all these mutations as covariates, only CALR (P<0.0001) and ASXL1 (P<0.0001) mutations remained significant as favorable and unfavorable prognostic markers, respectively.
In multivariable analysis, CALR mutations had a favorable impact on survival that was not accounted for by DIPSS-plus risk stratification (P=0.005), ASXL1 mutation status (P<0.0001) or both (P=0.0009); in the latter analysis that included both DIPSS-plus risk score and ASXL1 mutation status as covariates, the HRs (95% CI) were 3.5 (1.8–6.6) for triple-negative, 2.2 (1.4–3.5) for JAK2-mutated and 1.7 (0.9–3.3) for MPL-mutated cases. It is important to note that the multivariable analysis also identified higher DIPSS-plus risk score (P<0.0001) and mutant ASXL1 (P=0.005) as independent predictors of poor survival. Prognosis was particularly worse for triple-negative patients with median survival of only 2.5 years (Figure 2). Triple-negative status was also associated with inferior LFS (P=0.003; Figure 3). Finally, Figure 4 shows survival data stratified by CALR and ASXL1 mutation status and highlights the prognostically detrimental effect of ‘CALR–ASXL1+’ mutation status.
The current study confirms the frequent (74%) but not invariable presence of CALR mutations in JAK2/MPL-unmutated PMF.18, 19 In addition, the study provides detailed clinical, cytogenetic and molecular data in CALR-mutated PMF. The previously remarked association between mutant CALR and higher platelet count18, 19 was further elaborated in the current study by showing significantly higher incidence of thrombocytosis and lower incidence of thrombocytopenia in CALR-mutated patients, who were also less likely to display anemia or leukocytosis. The association with anemia appears to be in the opposite direction to what has been reported in ET18, 19 and might be partly related to the significantly lower frequency of U2AF1 mutations in CALR-mutated cases.25 Additional favorable prognostic features in CALR-mutated patients, described for the first time in the current study, included younger age and lower DIPSS-plus risk score. We also show significantly lower frequency of spliceosome mutations25, 28 in CALR-mutated cases and, otherwise, lack of other molecular or cytogenetic associations.
In terms of prognostic relevance, our observation regarding the superior prognosis of patients with CALR mutations, compared with those with JAK2 mutations, was similar to that of Klampfl et al.19 However, we provide additional novel information by showing that the prognostic relevance of CALR mutations in PMF was independent of age, DIPSS-plus risk score22 and ASXL1 mutation status.29 In the process, we identified ‘CALR-/ASXL1+’ mutation profile as being particularly detrimental to survival. CALR-mutated patients also fared better than triple-negative cases; the latter also displayed worse leukemia-free survival. Thus, ‘CALR-ASXL1+’ and ‘triple-negative’ PMF should now be considered as a molecularly high-risk disease, as is the case with patients with ASXL1,29 EZH2,30 SRSF226 or IDH27 mutations or unfavorable karyotype.21, 31, 32
The practical implications from the current study might be summarized as follows: (i) bone marrow morphology remains the cornerstone of PMF diagnosis because up to 26% of JAK2/MPL-negative cases still lack MPN-characteristic/specific molecular marker, (ii) morphological diagnosis of JAK2/MPL-negative PMF can now be complemented by CALR mutation screening, (iii) CALR mutations provide additional prognostic value in PMF, especially when combined with ASXL1 mutation status, with ‘CALR-ASXL1+’ being the most detrimental mutation profile and (iv) ‘triple-negative’ PMF should now be regarded as representing high risk disease, although higher number of patients should be studied in order to validate the observations from the current study.
About this article
A comparison of qPCR and ddPCR used for quantification of the JAK2 V617F allele burden in Ph negative MPNs
Annals of Hematology (2018)