Essential Thrombocythemia

Mutational Spectrum in Essential Thrombocythemia

The hypersensitivity of ET progenitor cells to a variety of cytokines is due to the clonal acquisition of driver mutations ( JAK2 , MPL , or CALR ) that activate the JAK-STAT signaling pathway, allowing hematopoiesis to occur in the absence of exogenous cytokines. The JAK2 V617F mutation occurs in 50% to 60% of ET patients, while recurrent CALR mutations occur in 15% to 24% of patients, and 3% to 5% have MPL mutations. The patients with ET who lack such driver mutations are said to be “triple negative.” While the driver mutations were originally thought to be mutually exclusive, it has recently been identified that a small subset of ET patients with low VAF JAK2 V617F can have co-occurring CALR or MPL mutations.

JAK2 is a cytoplasmic tyrosine kinase that plays a key role in mediating intracellular signaling from a variety of growth factors, including IL-3, erythropoietin, granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), and thrombopoietin. Co-expression of JAK2 V617F with a homodimeric type 1 cytokine receptor (including erythropoietin, thrombopoietin, or G-CSF) is necessary for hormone activation of JAK-STAT (signal transducer and activator of transcription) signaling pathways and for hematopoietic cell proliferation to become growth factor–independent. The JAK2 V617F mutation is present in ET patients with both clonal and polyclonal hematopoiesis. Patients with clonal hematopoiesis have a higher JAK2 V617F allele burden (26%) than patients with polyclonal hematopoiesis (16%). The relative size of the JAK2V617F clone is often small and remains stable over time in patients with both clonal and polyclonal ET. A VAF higher than 50% indicating the presence of granulocytes, homozygous for JAK2 V617F, has been found in 70% of PV patients, but occurs less frequently in ET. In ET patients, the increase in the JAK2 V617F VAF from a level well below 50% at presentation, to a level above 50% after years of follow-up could be indicative of the imminent evolution of the patient’s ET to PV. All PV patients have assayable erythroid colonies that are homozygous for JAK2 V617F, even in PV patients with a low burden of JAK2 V617F. By contrast, hematopoietic colonies cloned from ET patients are only occasionally JAK2 V617F homozygous colonies. The transition from JAK2 heterozygous to homozygous progenitors is a consequence of homologous recombination. These studies suggest that such an event is characteristic of PV but rarely occurs in ET. If there is a substantial increase in the numbers of JAK2 V617F homozygous progenitors in an ET patient, it would likely lead to a transition from an ET to a PV phenotype.

STATs are activated downstream of JAK2 V617F. STAT3 is a pivotal regulator of megakaryocytopoiesis, which might provide an explanation for its exclusive upregulation in ET. To further examine the differences between hematopoiesis in JAK2 V617F ET and PV, the gene expression profiles of JAK2 V617F-heterozygous erythroid colonies have been examined. Erythroblasts from ET patients were characterized by enhanced expression of genes associated with interferon (IFN) signaling and phosho-STAT1, compared with PV erythroblasts. STAT1 is essential for IFN-γ signaling. Increased STAT1 in normal CD34 ⁺ cells has been shown to favor megakaryocytic differentiation but reduce erythroid differentiation, creating a differentiation pattern that resembles ET. Furthermore, inhibition of STAT1 signaling in ET hematopoietic progenitor cells led to enhanced erythropoiesis and reduced megakaryocytopoiesis. These studies suggest that in ET, JAK2V 617F induces the simultaneous activation of STAT5 and STAT1 pathways, but in PV, the relatively reduced levels of phospho-STAT1 reduce the brake on erythropoiesis. Further functional annotation analyses have also demonstrated that clusters of gene ontology terms related to inflammation, immune system, apoptosis, RNA metabolism, and secretory system are also significantly dysregulated terms in ET, PV, and MF. These data support the concept of MPNs being a biological continuum transitioning from ET to PV to PMF.

Thrombopoietin is the primary physiologic regulator of thrombopoiesis, which acts by binding to its cell surface receptor, MPL. Thrombopoietin levels in a normal individual are inversely related to megakaryocyte mass. MPL is expressed by CD34 ⁺ hematopoietic stem and progenitor cells, MKs, and platelets. One would anticipate that thrombopoietin levels in ET would be low due to the characteristic megakaryocytic hyperplasia. Normal or slightly elevated thrombopoietin levels have, however, been observed in patients with ET. Furthermore, Wang et al. have reported that thrombopoietin levels were significantly higher in patients with ET than in patients with reactive thrombocytosis who had similar levels as control. This unanticipated elevation of thrombopoietin levels in ET patients might be explained by the observation that the expression of the thrombopoietin receptor and its mRNA is dramatically reduced in the platelets of patients with ET. Thrombopoietin serum levels are controlled by platelet mass through MPL-mediated thrombopoietin uptake and degradation. The reduced platelet MPL expression occurs not only in ET but also in PV and PMF and has been shown to be a downstream event of JAK2 V617F, which promotes the proteasomal degradation of MPL. The reduced MPL likely results in the decreased capacity of platelets to absorb thrombopoietin, contributing to the increased megakaryocyte mass and thrombocytosis. The mutations in the thrombopoietin receptor, MPL W515L and MPL 515K, are present in approximately 3% to 5% of patients with ET, and additional novel MPL mutations have been identified in ET patients in a recent study utilizing deep mutational scanning of MPNs and will require further investigation. About 60% of patients with MPL mutations have the W515L mutation and 40% have the W515K mutation. The mutant allele burden is greater than 50% in 50% of W515K patients compared with 17% of W515L patients. The most prevalent MPL mutations in ET occur on tryptophan 515, an amino acid that maintains MPL in an inactive form in the absence of cytokines. Rarely, in ET patients, another MPL mutation, S505N, serves as a driver mutation. It is located in the exon 10 domain that encodes the transmembrane domain of MPL and induces dimerization of the transmembrane helix in an active confirmation. The MPL mutations occurring in ET trigger conformational changes in the receptor, bringing into close proximity two molecules of bound JAK2, for transphosphorylation and activation of the JAK-STAT signaling cascade. The loss of tryptophan, but not the acquisition of a particular residue, induces the constitutive activation of MPL. In ET, both JAK2 V617F and MPL mutations arise preferentially on a specific constitutional JAK2 46/1 haplotype. Two hypotheses have been proposed to account for this predilection: 46/1 is inherently genetically more unstable (hypermutability hypothesis) or 46/1 confers a growth advantage that favors the predominance of JAK2 V617F hematopoiesis (fertile ground hypothesis). The association of MPL mutations with the JAK2 46/1 haplotype strongly favors the hypermutability hypothesis rather than the fertile ground hypothesis. The presence of MPL W515K mutations in ET patients is associated with lower hemoglobin levels and higher platelet counts, as well as preferential expansion of the numbers of megakaryocytes at the expense of erythroid precursors, as observed in BM biopsy specimens.

Of patients with ET or MF who do not have mutations in JAK2 or MPL , the overwhelming majority have a mutation in CALR . Mutations in the CALR gene occur in 15% to 24% of ET patients and rarely occur together with mutations of JAK2 or MPL . The clinical phenotype of patients with JAK2 -mutated and CALR -mutated ET differs. The presence of the mutated CALR is associated with a younger age, male predominance, higher platelet counts, lower hemoglobin levels, lower leukocyte counts, and a lower risk of thrombotic complications. Recent studies have suggested that the lower thrombotic rates in CALR -mutated ET may in part be due to reduced leukocyte activation or less reactive platelets, compared with their JAK2 V617F-positive counterparts. CALR -mutated ET patients do not evolve to PV while this transition occurs in 20% to 30% of JAK2 V617F+ ET patients over a 15-year period. Grinfeld et al. have independently reported that CALR mutations are independently associated with an increased risk of developing MF.

The wild-type CALR gene encodes for an evolutionarily conserved, multifunctional protein, involved in multiple cellular processes ranging from calcium homeostasis and protein folding in the endoplasmic reticulum, to apoptotic cell death clearance and cellular adhesion. The CALR mutations ( CALR : 19p13.2) identified in MPN mainly consist of deletions (i.e., type I) or insertions (i.e., type II) occurring within exon 9, which create a novel epitope in the C-terminal domain of the protein. Despite the heterogeneity of these mutations, the new C-terminus sequence is identical and results in the loss of the KDEL domain, which is critical for CALR retention in the endoplasmic reticulum and its ability to regulate calcium homeostasis. Original studies describing CALR mutations in MPN indicated that such a uniform defect may confer a proliferative advantage to malignant cells via activation of the JAK-STAT pathway.

Mutations of CALR are very rare in PV but are found in patients with MF and ET, the two MPN entities in which megakaryocytic hyperplasia is a hallmark of the disease. The most frequent CALR mutations, accounting for more than 80% of the total, are the type-1 variant, a 52 bp deletion (p.L367fs*46), and the type-2, a 5 bp TTGTC insertion (p.K385fs*47). The other types of mutations have been observed at much lower frequencies, and many have been detected only in individual patients. Overall, type-1 is more frequent (53%) than type 2 mutations (31.7%), but the incidence of type-2 mutations is higher in ET as compared to MF. Type 1 mutations are more frequent in MF patients. Some investigators have conjectured that type 1 CALR mutations are a forme fruste of MF. Most CALR mutations are present in a heterozygous state, although rare patients are homozygous for CALR mutations, which are associated with uniparental disomy of chromosome 19p. The CALR mutations target the hematopoietic stem cell and the mutation can be detected in granulocytes, monocytes, platelets, B, and NK cells. Occasionally in ET patients, they may be seen in T cells. Both types of CALR mutations lead to an increase in the number of megakaryocyte progenitors that are hypersensitive to cytokines. A series of subsequent reports have shed light on the mechanisms responsible for the ability of the mutated CALR to induce MK hyperplasia and thrombocytosis in MPN. Thus four independent laboratories have demonstrated that the oncogenic activity of mutated CALR is mediated by MPL, which is critical for both hematopoietic stem cell and megakaryocytic lineage development. Thus mutant CALR expressed by hematopoietic cells in vitro or in vivo in mouse models binds and activates MPL but not other type I hematopoietic cytokine receptors. This, in turn, leads to JAK-STAT pathway activation, resulting in the development of thrombocytosis and an ET-like phenotype. Interestingly, in vivo overexpression of either of the mutation variants (type I, deletion; or type II, insertion) induces thrombocytosis and constitutively activates MPL and JAK-STAT signaling. Yet, while type II mutations favor MK proliferation, type I mutations confer clonal dominance of the hematopoietic stem cell, resulting in splenomegaly, BM hypocellularity, and fibrosis, a phenotype reminiscent of “post-ET” MF. These findings mirror the clinical observations in which type I mutations are more prevalent in MF patients, while type II mutations are more frequent in ET patients. Moreover, CD34 ⁺ hematopoietic progenitors from ET patients harboring the type II CALR mutation are capable of forming spontaneous CFU-MK, validating the initial suggestions that the mutated CALR confers cytokine-independent cell growth.

Elegant molecular and biochemical studies dissecting the interaction between mutant CALR and MPL and its downstream consequences revealed that the N-glycosylation sites of the MPL extracellular domain are required for its activation by mutated CALR and are independent of the thrombopoietin binding site. Furthermore, the mutated CALR translocates to the cell surface and acts in an autocrine manner by binding and activating MPL. Recently, it has been postulated that the interaction between mutant CALR and MPL may occur in the endoplasmic reticulum. Although the exact mechanisms by which the mutated CALR binds to the MPL are under investigation, preliminary evidence suggests that the oncogenic properties of the mutated CALR can be attributed to the positive electrostatic charge of the novel peptide, which may be responsible not only for the physical interaction with MPL but also for other cellular functions involving the normally negatively charged portion of the protein ( Fig. 71.1 ). Mutated CALR can also affect the immune system. When CALR is present on the cell surface, it acts as an “eat me” signal that facilitates the recognition of stressed cells by phagocytes that express the CALR receptor, CD91. This type of innate immune response facilitates the transfer of tumor antigens to dendritic cells, thereby facilitating the induction of cancer-specific cytotoxic T cells. The mutated CALR proteins escape from the ER system and can be secreted into the tissue microenvironment. Liu et al. have reported that soluble CALR proteins can, in fact, act as decoy receptors, preventing the uptake of CALR-expressing cells by dendritic cells, thereby leading to profound immunosuppression. CALR mutations may therefore play two roles in ET and MF (1) activating cell autonomous oncogenic drivers and (2) blunting the immune response. In fact, immune suppressive mechanisms, including T-cell exhaustion, have been reported in patients harboring mutant CALR , perhaps due to its effects on the dendritic cell activation reported above. This immune suppressive effect of CALR-mutated proteins has been proposed as the reason why CALR- mutated MPNs arise 10 years earlier than JAK2 V617F+ MPNS. Mutant CALR has recently been proposed as an MPN-specific neoantigen that might serve as a target for immunotherapeutic approaches. CALR-specific T-cell responses have been partially restored in vitro and in vivo by antibodies directed against immune checkpoint inhibitors. The success of such vaccine strategies, however, might require the development of strategies directed towards restoring immune responses in patients by overcoming the effects of secreted mutated CALR.

The SH2B adaptor protein 3 (SH2B3) gene, also known as the LNK gene, encodes a negative regulator of cytokine signaling. Occasional cases of ET are associated with loss-of-function Lnk (SH2B3) mutations. In mouse models, the inhibitory adaptor protein, LNK, has been shown to be associated with the downmodulation of erythropoietin and thrombopoietin signaling. Lnk can bind to wild-type JAK2 , JAK2 V617F, wild-type MPL , and MPL W515L. Lnk levels are upregulated and they correlate with an increase in the JAK2 V617F VAF in MPN patients. In JAK2 V617F-positive ET, Lnk mRNA expression is upregulated and serves to modulate JAK2 V617F-mediated cell regulation. Thrombopoietin-mediated signaling regulates Lnk expression at both the mRNA and protein levels. Furthermore, acquired Lnk mutations have been observed in less than 1% of ET. The inactivating Lnk mutations in ET patients result in JAK-STAT activation, leading to high levels of STAT3 and STAT5 activation. Cabagnois and coworkers used whole-exome sequencing and next-generation sequencing, targeting JAK2 and MPL , with the intent of detecting additional mutations in triple negative ET patients. They found several signaling mutations, including JAK2 V617F with very low VAFs, as well as additional mutations such as: LNK mutation, MPL- S505N, MPL-W515R, and MPL- S204P. MPL- S204P and MPL- Y591N were shown to be weak gain-of-function mutants, increasing MPL signaling and conferring either thrombopoietin hypersensitivity or independence to expressing cells, but with low efficiency. These data demonstrate that some clonal, noncanonical MPL gain-of-function mutations are associated with triple negative cases of ET. “Triple negative” ET patients have also been shown to harbor clonal markers involving other myeloid associated genes, including ten–eleven translocation ( TET2 ) and KIT . In addition, a recent study showed that nearly a quarter of those ET patients diagnosed as “triple negative” actually harbored low-level JAK2 or MPL mutations, detected using deep next-generation sequencing. The study also revealed that “true triple negative” ET patients and ET patients with known driver mutations had similar gene expression profiles consistent with activation of the JAK/STAT and nuclear factor kappa B (NFkB) pathways. Ultimately, patients with triple negative ET cannot always be classified as having an MPN and should be evaluated for inherited forms of thrombocytosis.

Mutations in epigenetic regulators (such as TET2 , DNMT3A , ASXL1 , EZH2 , and isocitrate dehydrogenase [ IDH ] 1/IDH2 ) and in spliceosome components (such as SRSF2 , U2AF1 , and SF3B1 ) have been shown to be present in MPN patients with JAK2 / MPL / CALR mutations. Other mutations were also directly associated with leukemic progression, such as TP53 , RUNX1 , CBL , and deletion in IKAROS . Several of these mutations can occur in the same patient, and most frequently SRSF2 can occur with TET2 , ASXL1 , or IDH mutations. In contrast to MPN driver mutations, which are rare in other myeloid malignancies, these additional mutations are not specific to MPNs and are found with a higher frequency in patients with myelodysplastic syndrome (MDS) and in MDS/MPN overlap syndromes, such as chronic myelomonocytic leukemia (see Chapter 73 ) as well as AML. Biologic studies and mouse models have shown that these additional mutations may cooperate with MPN driver mutations to favor clonal dominance ( TET2 or DNMT3A ), to modify disease phenotype, or to promote either progression to MF or leukemic transformation ( ASXL1 , IDH1/2 , EZH2 , and TP53 ). Tefferi and coworkers have reported that the percentage of ET patients with 1, 2, or ≥3 sequence variants/mutations was 41%, 8%, and 4%, respectively, with the most frequent sequence variants/mutations in TET2 and ASXL1 . In addition, a significant association between the MPN driver mutational status and the number or type of other sequence variants/mutations in the Mayo Clinic series of 183 patients with ET was not observed. Especially important was their stratification of survival by the presence or absence of “adverse” ( SH2B3 , IDH2 , SF3B1 , U2AF1 , EZH2 , TP53 ) or “other” ( TET2 , ASXL1 , PTP11 , SUZ12 , ZRSR2 , CBL , CEBPA , CSF3R , DNMT3A , SRSF2 , FLT3 , KIT , NRAS , RUNX1 , SETBP1 ) sequence variants/mutations. Those patients in a cohort of 174 Italian patients with ET, with adverse mutations, had a median survival of 18.4 years as compared to 29 years for those individuals who harbored other vatiants that were not high-risk. A recent study evaluating the genomic profiles of ET and PV patients who undergo leukemic transformation revealed that DNMT3A , U2AF1 , IDH1 , IDH2 , and EZH2 mutations were associated with early transformation (median time to transformation 3 years) and TP53 , BCORL1 , and NRAS mutations were associated with late transformation (median time to transformation 21 years).

The TET2 gene encodes a hydroxylase that is able to hydroxylate methylated cytosine. These mutations result in a loss of function, leading to increased DNA methylation and a reduction in hydroxymethylcytosine. TET2 mutations occur in 11% of ET patients. Mutations in TET2 can occur either before or after mutations of JAK2 or MPL . Mutations in the gene for IDH occur in 0.9% of ET patients and can functionally lead to similar effects as TET2 mutations on DNA methylation. Furthermore, disrupting mutations of ASXL1 , which occur in 36% of PMF patients, are infrequent in ET, as are mutations of CBL or EZH2 .

Furthermore, the transcription factor NFE2 has been shown to be overexpressed in the cells of patients with MPNs independent of the presence or absence of JAK2 V617F. The NFE2 acts as an epigenetic transcriptional regulator and a chromatin modifier. It is part of a complex regulatory network, including transcription factors such as GATA1 and RUNX1, controlling megakaryocytic and/or erythroid cell function. The knockout mouse model displays only mild erythroid abnormalities, while the major phenotype is a defect in megakaryocyte biogenesis. Indeed, the absence of NFE2 leads to severely impaired platelet production. Genetically engineered mice that overexpress NF-E2 are characterized by extreme thrombocytosis and leukocytosis, normal hemoglobin levels, and BM hypercellularity, which is a clinical picture similar to that observed in ET patients. Frameshift and deletion NFE2 mutations are found in approximately 2% to 4% of ET patients and have been shown to be preferentially associated with MPL (30%) and CALR (40%) mutations. While there was no direct correlation between NF-E2 and JAK2 V617F VAF in different MPNs, the frequency of NFE2 mutations resulting in a functionally enhanced truncated form of NF-E2 was doubled in patients with greater than 50% VAF (9.7% vs. 4.1%). Furthermore, Grinfeld and coworkers reported that the presence of NF-E2 mutations was correlated with the transition of JAK2 V617F+ ET patients to a PV phenotype. Mice expressing NF-E2 mutations develop erythrocytosis and thrombocytosis due to the expansion of their progenitor and stem cell compartments, and in the long term may develop leukemias with concomitant and/or isolated myelosarcoma. Marcault and coworkers reported that MPN patients with NF-E2 mutations with VAFs greater than 5% were at an increased risk of transformation to MPN-BP.