Calreticulin: a new horizon for the testing and treatment of myeloproliferative neoplasms

Abstract

The recent discovery of mutations of the gene calreticulin has allowed raising the proportion of patients with essential thrombocythemia and primary myelofibrosis with known mutational abnormality up to 85–90%. Knowledge of the mechanisms by which mutated calreticulin underlie a myeloproliferative neoplasm as well as the clinical and therapeutic implications is just at the very beginning, and exciting times await research in this field.

Keywords::

• calreticulin
• essential thrombocythemia
• JAK/STAT signaling
• myeloproliferative neoplasms
• primary myelofibrosis

The Philadelphia-chromosome-negative classic chronic myeloproliferative neoplasms (MPNs), which include polycythemia vera (PV), essential thrombocythemia (ET) and primary myelofibrosis (PMF), currently represent one of the most active and exciting fields in experimental and clinical hematology. After decades of sluggish advancements, a burst of knowledge was primed by the 2005’ four reports describing a recurrent JAK2V6717F mutation [1]. This mutation resulted associated with the majority of cases with a PV phenotype (≥95%) and about 60% of ET and PMF; a 2–3% of PV patients wild-type for the V617F allele harbor different JAK2 mutations located in exon 12, while 3–8% of ET and PMF patients present mutations in the gene encoding for the thrombopoietin receptor, MPL [1].Those discoveries prompted the WHO panelists to revise the diagnostic criteria of MPN in 2008 [2], oriented basic researchers toward the central pathogenetic involvement of the JAK/STAT signaling pathway [3] and finally led, in an unusually short lapse of time, to the completion of two Phase III trials and approval of the first JAK2 (and JAK1) inhibitor for the treatment of patients with myelofibrosis [4,5]. However, several other aspects remain at present largely unsolved, which include whether JAK2V617F is the founding mutation in these disorders, what is the mechanistic role of additional somatic variations that were discovered later in genes of the epigenetic machinery and the spliceasome and how the burden of mutated alleles influences phenotypic presentation and prognosis. Therefore, while virtually all patients with PV resulted characterized by a mutation in JAK2, >30% of ET and PMF patients lacked a recurrent molecular marker; this situation changed suddenly in December 2013 when J. Nangalia from the group of Tony Green in Cambridge [6] and T. Klapmfl from Robert Kralovics’ laboratory in Wien [7] concurrently reported the discovery of novel mutations in the gene calreticulin (CALR). Mutations in CALR were found in 60–88% of patients with ET and PMF who were JAK2 and MPL unmutated, with the proportion of ET and PMF patients that remained molecularly uncharacterized (‘triple negative’) falling down to <15%. CALR mutations are represented by insertions or deletions restricted to exon 9; 80–90% of mutated cases harbor either a 52–bp deletion (Type I; 45–53% of all cases) or a 5-bp insertion (Type II; 32–41%). Remarkably, all mutations cause a frameshift, resulting in a novel C-terminal peptide made up by a minimal 36-amino acid stretch in place of 27 amino acids that are lost from the normal sequence.

Calreticulin is a multicompartmental and multifunctional protein whose best known function is that of a Ca²⁺-binding chaperone in the endoplasmic reticulum (ER) lumen [8]. Disruption of CALR is embryonically lethal due to defects in heart development attributable to impaired myofibrillogenesis because of an abnormal ER calcium availability [9]. The mature protein form consists of three distinct domains: a globular N-terminal domain containing a signal sequence for ER targeting, a middle P-domain that contains high-affinity, low-capacity, binding sites for Ca²⁺ and a highly acidic C-terminal domain that has Ca²⁺-binding properties with high capacity and low affinity. The terminal part of the C-domain contains a KDEL sequence, which is normally involved in the retrieval of KDEL-containing proteins from the cis-Golgi back to the ER. In concert with other ER-associated proteins, particularly calnexin, CALR intervenes in ensuring proper protein and glycoprotein folding. Calreticulin has structural homology with calnexin, and both function in concert in the ‘calnexin/calreticulin cycle’, an N-glycan-dependent quality control process that ensures correct folding and/or degradation of glycoproteins and prevents protein aggregation [10].

Based on what was known at that time, mutated calreticulin was a totally unexpected new player in the arena of MPN, and how it is involved in the development of a myeloproliferative neoplasia remains largely unknown; as a matter of fact, there is little information concerning the function played by calreticulin in normal hematopoiesis. However, it is possible that the key pathogenetic effects attributable to mutated calreticulin pertain, at least in part, to the JAK/STAT signaling, as supported by findings that expression of CALRdel52 in interleukin-3-dependent murine Ba/F3 cells was associated with increased STAT5 phosphorylation and caused the cells to become cytokine independent; the growth of CALR mutated cells resulted inhibited by a JAK2 inhibitor [7].These in vitro observations match well with reported effectiveness of JAK2 inhibitor therapy in CALR mutated patients [11]. Due to the multiplicity of actions normally performed by calreticulin in the cell, it is very likely, however, that mechanisms leading to over-activation of JAK/STAT signaling are complex and only partially reflect the derangement of cellular functions generated by the mutated protein. The C-terminus of mutated CALR lacks KDEL motif, which raises the possibility of cellular mislocalization with consequently altered functions; preliminary studies in cells overexpressing mutated CALR, as well as in fresh myeloid cells from CALR mutated patients, have produced conflicting results, although it appears that mutant CALR may largely retain its localization within the ER, in line with the presence of KDEL-independent mechanisms for ER retention associated with the acidic region of calreticulin [10]. A bioinformatic analysis of the structure–function relationships of mutated CALR indeed suggests that the C-terminus of mutated protein loses the capacity to bind Ca²⁺, which in turn may affect a variety of cellular functions [12]. Deficiencies in calreticulin/calnexin may be involved in altered lineage commitment of hematopoietic progenitors, resulting in a preferential expansion of the megakaryocytic cell lineage [13]. Beyond the ER, wild-type calreticulin was reported to localize in the cytosol, the nucleus, at the cell surface and also in extracellular fluids. Of interest, leukemic blasts and dysplastic progenitors from patients with myelodysplastic syndromes showed upregulation of CALR on the cell surface; calreticulin expressed on the cell membrane might represent an ‘eat me’ signal, eventually counterbalanced by expression of CD47, thereby affecting the phagocytosis of neoplastic cells [14]. However, no significant increase in CALR content at the cell surface was found in CALR mutated granulocytes [6].

It is anticipated that discovery of CALR mutations will impact significantly the way we approach MPN in the different fields of experimental research, disease modeling, patients’ diagnosis and management and, prospectively, therapy. JAK2 and MPL mutations constituted major diagnostic criteria in the revised 2008 WHO classification of MPN [2]. Outside JAK2/MPL wild-type ET and PMF patients, CALR mutations were found very infrequently in a few cases of atypical chronic myeloid leukemia or chronic myelomonocytic leukemia and in occasional patients with myelodysplastic syndromes, particularly with refractory anemia with ring sideroblasts associated with marked thrombocytosis. Lymphoid neoplasia, acute myeloid leukemias and solid tumors resulted unmutated [6,7]. Therefore, due to the frequency and selective association of CALR mutations with ET and PMF, it was proposed that CALR mutations might be endorsed by the WHO committee for inclusion among the major diagnostic criteria for those disorders side by side with JAK2 and MPL mutations [15], thereby facilitating discrimination of non-clonal, reactive forms of thrombocytosis and/or myelofibrosis. However, bone marrow biopsy remains mandatory for the differential diagnosis between ET and PMF, including prefibrotic myelofibrosis, as well as for the diagnosis of the ‘triple-negative’ patients. Keeping this in mind, we suggested that whenever genotyping for CALR mutations is not available, immunolabeling with an antibody specifically directed against mutated calreticulin antibody could serve as a tool to identify subjects harboring CALR mutations [16]. A CALR mutated genotype seems to deserve relevance also for prognostic stratification. In ET, CALR mutated patients are at significantly lower risk of thrombosis [7,17,18] while overall survival of CALR mutated PMF patients was significantly better compared to other genotypes, particularly the ‘triple-negative’ patients [19], and especially in the comparison with patients concurrently negative for CALR and positive for the prognostically detrimental ASXL1 mutations [20]. More subtle phenotypic effects may also be dependent on the unique molecular lesion in CALR, with Type 1 mutation being reported to account for a better prognosis than Type 2 in patients with PMF [21]. It is still premature to assume that mutated calreticulin might represent a new therapeutic target until a more definite picture of the molecular basis of CALR mutated diseases can eventually be obtained. However, based on the preliminary information we have, and retrospectively on the observation that no significant difference was noted in the clinical response to JAK2 inhibitors between JAK2V617F mutated and unmutated PMF patients (the large majority of which we now know were harboring CALR mutation), it is reassuring that inhibition of the abnormal JAK/STAT signaling represents a valuable therapeutic option also for CALR mutated patients. Whether it will be feasible to develop specific and more effective inhibitors of mutated calreticulin remains a formidable challenge for the near future.

Financial & competing interests disclosure

This study was supported by a special grant from Associazione Italiana per la Ricerca sul Cancro-“AIRC 5 per Mille” – to AGIMM, “AIRC-Gruppo Italiano Malattie Mieloproliferative” (#1005); for a description of the AGIMM project and list of investigators, see at www.progettoagimm.it). Partially supported by Ministero della Università e Ricerca (MIUR; FIRB project #RBAP11CZLK and PRIN 2010NYKNS7). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.