Uses and Abuses of JAK2 and MPL Mutation Tests in Myeloproliferative Neoplasms

Abstract

JAK2V617F is sufficiently prevalent in BCR-ABL1-negative myeloproliferative neoplasms (MPNs) to be useful as a clonal marker. JAK2V617F mutation screening is indicated for the evaluation of erythrocytosis, thrombocytosis, splanchnic vein thrombosis, and otherwise unexplained BCR-ABL1-negative granulocytosis. However, the mutation does not provide additional value in the presence of unequivocal morphologic diagnosis, and its presence does not necessarily distinguish one MPN from another or provide useful prognostic information. In general, quantitative cell-based JAK2V617F mutation assays are preferred because the additional information obtained on mutant allele burden enhances diagnostic certainty and facilitates monitoring of response to treatment. JAK2 exon 12 mutation screening is indicated only in the presence of JAK2V617F-negative erythrocytosis that is associated with a subnormal serum erythropoietin level. MPL mutations are neither frequent nor specific enough to warrant their routine use for MPN diagnosis, but they may be useful in resolving specific diagnostic problems. The practice of en bloc screening for JAK2V617F, JAK2 exon 12, and MPL mutations is scientifically irrational and economically irresponsible.

Morphology is the cornerstone of current diagnosis and classification in myeloid malignancies.¹ Cytochemical, immunophenotypic, cytogenetic, and molecular data enhance diagnostic accuracy and form the basis for the World Health Organization classification of myeloid malignancies into five main categories²: acute myeloid leukemia, myelodysplastic syndromes (MDSs), myeloproliferative neoplasms (MPNs), MDS/MPN overlap, and platelet-derived growth factor receptor gene or fibroblast growth factor receptor 1 gene rearranged myeloid/lymphoid neoplasms associated with eosinophilia. The World Health Organization MPN category includes eight subcategories: chronic myelogenous leukemia, polycythemia vera (PV), essential thrombocythemia (ET), primary myelofibrosis (PMF), mastocytosis, chronic eosinophilic leukemia not otherwise specified, chronic neutrophilic leukemia, and MPN unclassifiable.¹ Among these subcategories, the first four (ie, chronic myelogenous leukemia, PV, ET, and PMF) are currently referred to as “classic” MPNs, because they were included in the original description of myeloproliferative disorders by William Dameshek.³ In general, current evidence supports consideration of all myeloid malignancies, including MPN, as clonal stem cell diseases.

JAK2 and MPL mutations occur across the spectrum of myeloid malignancies, including MPN, MDS, MDS/MPN, and acute myeloid leukemia (Table 1).^4–7 These mutations are most prevalent in the BCR-ABL1-negative classic MPN (ie, PV, ET, and PMF), which are morphologically characterized by the absence of both cellular dysplasia and monocytosis and the presence of abnormal megakaryocytes that are increased in number and often found in clusters. Megakaryocyte morphology and degree of trilineage proliferation differ among the three BCR-ABL1-negative classic MPNs. Megakaryocytes are large, hyperlobulated, and mature appearing in ET⁸; immature appearing with hyperchromatic and irregularly folded bulky nuclei in PMF⁸; and pleomorphic without maturation defects in PV.^9,10 Megakaryocyte changes in BCR-ABL1-negative MPN are accompanied by left-shifted granulocyte proliferation in PMF, trilineage proliferation in PV, and otherwise normal-appearing bone marrow in ET. Overt bone marrow fibrosis is absent in prefibrotic PMF, which is otherwise characterized by the aforementioned PMF-associated changes in megakaryocyte morphology and increased granulocyte proliferation.¹¹ Controversy is ongoing about the utility of morphology alone to distinguish ET from early PMF¹²; however, this is irrelevant because clinical pathologists never base their diagnostic impressions on morphology alone, and they also consider clinical, cytogenetic, and molecular information.^1,11

Table 1.

Currently Known Mutations in BCR-ABL1-Negative Myeloproliferative Neoplasms

Mutations	Chromosome location	Mutational frequency, %
JAK2	9p24
PV		∼964
ET		∼554
PMF		∼654
BP-MPN		∼504
JAK2 exon 12 mutation [4]	9p24
PV		∼34
MPL	1p34
ET		∼34
PMF		∼104
BP-MPN		∼54
LNK	12q24.12
PV		Rare20,21
ET		Rare19,20
PMF		Rare19,20
BP-MPN		∼1020
TET2	4q24
PV		∼164
ET		∼54
PMF		∼174
BP-MPN		∼174
ASXL1	20q11.1
ET		∼332
PMF		∼13
BP-MPN		∼18
IDH1/IDH2	2q33.3/15q26.1
PV		∼224
ET		∼124
PMF		∼424
BP-MPN		∼2024
EZH2	7q36.1
PV		∼334
PMF		∼7
DNMT3A	2p23
PV		∼735
PMF		∼735,36
BP-MPN		∼1435,36
CBL	11q23.3
PV		Rare22
ET		Rare22
MF		∼622
IKZF1	7p12
CP-MPN		Rare27
BP-MPN		∼1927

BP-MPN, blast-phase MPN; CP-MPN, chronic phase MPN; MF, both PMF and post-ET/PV myelofibrosis.

Clinically, PV and ET are characterized by erythrocytosis and thrombocytosis, respectively, and leukocytosis, splenomegaly, thrombohemorrhagic complications, vasomotor disturbances, pruritus, and a small risk of disease progression into acute leukemia or myelofibrosis.¹³ PMF is characterized by anemia, splenomegaly, extramedullary hematopoiesis, constitutional symptoms, and a higher risk of leukemic progression.¹⁴ Median survival exceeds 15 years in both ET and PV, but it is significantly shorter in PMF. The goal of therapy in PV and ET is to prevent thrombotic complications. Low-dose aspirin is the cornerstone of therapy in both PV and ET.¹⁵ In addition, phlebotomy is required in PV and hydroxyurea therapy in high-risk disease (history of thrombosis or age >60 years).¹³ PMF is managed according to risk category from the Dynamic International Prognostic Scoring System.¹⁶ Patients with low-risk or intermediate-1 PMF are managed by observation alone or conventional drug therapy, whereas allogeneic stem cell transplantation or experimental drug therapy might be necessary for intermediate-2 or high-risk disease.

Overview of Mutations Associated with BCR-ABL1-Negative MPNs

The disease-initiating mutation(s) in BCR-ABL1-negative MPN is unknown. However, JAK2V617F is present in most patients with PV, ET, or PMF, and a minority of patients with these diseases also harbor JAK2 exon 12, MPL, LNK, CBL, TET2, ASXL1, IDH, IKZF1, EZH2, or DNMT3A mutations (Table 1).⁴ These mutations are currently thought to represent secondary events and to lack both disease specificity and mutual exclusivity. Some patients carry more than one mutation, and clonal hierarchy in such instances appears to be unpredictable.¹⁷

JAK2 (Janus kinase 2) maps to chromosome 9p24. JAK2V617F is located on exon 14 and occurs in ∼96% of patients with PV, 55% with ET, and 65% with PMF.⁴ JAK2V617F contributes to abnormal myeloproliferation in MPN, whereas such effect is erythroid lineage weighted with JAK2 exon 12 mutation.⁴ MPL (myeloproliferative leukemia virus oncogene) maps to chromosome 1p34, and MPL mutations usually involve exon 10 and contribute to primarily megakaryocytic myeloproliferation.⁴ MPL mutational frequencies are estimated at 3% in ET and 10% in PMF.⁴ LNK (as in Links) maps to chromosome 12q24.12 and encodes for a membrane-bound adaptor protein that negatively regulates JAK2 signaling.¹⁸ LNK mutations usually involve exon 2, are inactivating, and occur in ∼10% of patients with blast-phase MPN, whereas they are infrequent in chronic-phase disease.^19–21

CBL (Casitas B-lineage lymphoma proto-oncogene) maps to chromosome 11q23.3, and its mutations involve exons 8 and 9.²² CBL is an E3-ubiquitin ligase that marks mutant kinases for degradation.²³ CBL mutations are rare in PV and ET but are reported to occur in ∼6% of patients in myelofibrosis.²² IDH1 and IDH2 (isocitrate dehydrogenase) map to chromosomes 2q33.3 and 15q26.1, respectively, and their mutations involve exon 4.²⁴ IDH mutations induce formation of 2-hydroxyglutarate, which is thought to be oncogenic.²⁵ IDH mutational frequencies are estimated at 2% in PV, 1% in ET, 4% in PMF, and 20% in blast-phase MPN.²⁴ IKAROS family zinc finger 1 (IKZF1) maps to chromosome 7p12 and is thought to function as a transcription regulator and tumor suppressor.²⁶ IKZF1 mutations are rare in chronic-phase MPN but might be detected in ∼19% of patients with blast-phase MPN.²⁷

TET oncogene family member 2 (TET2) maps to chromosome 4q24. TET2 mutations occur across several of the gene's 12 exons.⁴ TET proteins catalyze conversion of 5-methylcytosine to 5-hydroxymethylcytosine,^28,29 and, accordingly, TET2 mutations are thought to contribute to epigenetic dysregulation of transcription. TET2 mutational frequencies are estimated at 16% in PV, 5% in ET, 17% in PMF, and 17% in blast-phase MPN.⁴ Additional sex combs-like 1 (ASXL1) maps to chromosome 20q11.1, and ASXL1 mutations involve exon 12. Wild-type ASXL1 is needed for normal hematopoiesis³⁰ and might be involved in transcriptional repression.³¹ ASXL1 mutations are rare in PV or ET³² but were reported at ∼13% in PMF and 18% in blast-phase MPN.³³ Enhancer of zeste homolog 2 (EZH2) maps to chromosome 7q36.1, and EZH2 mutations involve exons 10, 18, and 20.³⁴ Wild-type EZH2 is part of a histone methyltransferase and might function both as a tumor suppressor and an oncogene.³⁴ EZH2 mutational frequencies are reported at ∼3% in PV³⁴ and 7% in PMF.³³

Most recently, DNA methyltransferase 3a (DNMT3A) mutations were reported in patients with acute myeloid leukemia, MDS, and MPN; mutational frequencies in the latter were reported at ∼7% for PV, 6% for PMF, and 14% for blast-phase MPN.³⁵ Another study of 46 patients with PMF, 22 with post-ET/PV myelofibrosis, and 11 with blast-phase MPN reported corresponding DNMT3A mutational frequencies of 7%, 0%, and 0%, respectively.³⁶ All 13 DNMT3A mutations reported in those two studies were heterozygous, and the most frequent mutations affected amino acid R882. DNMT3A mutations in those two MPN studies were documented to occur in the presence or absence of JAK2, IDH, ASXL1, or TET2 mutations.

In summary, activating JAK2 and MPL mutations and LNK loss-of-function result in constitutive JAK-STAT activation and induce MPN-like disease in mice.^19,37–39 TET2, ASXL1, and EZH2 mutations might contribute to epigenetic dysregulation of transcription.^28,29,31

JAK2 and MPL Mutation Screening in Routine Clinical Practice

The decision to screen for JAK2 or MPL mutations, in the context of routine clinical practice, is based on a number of facts and possibilities:

• •JAK2V617F is present in most patients with PV, ET, or PMF but lacks disease specificity or prognostic value.⁴⁰
• •JAK2V617F also occurs in other myeloid malignancies and is therefore useful as a clonal marker in the evaluation of otherwise unexplained BCR-ABL1-negative granulocytosis or monocytosis.⁴¹
• •JAK2V617F, but not JAK2 exon 12 or MPL mutations,⁴² has been shown to identify occult MPN in patients with splanchnic vein thrombosis,⁴³ but the yield of mutation screening in the evaluation of non-splanchnic thrombosis is very low.^44,45
• •JAK2 mutations are present in virtually all patients with PV; therefore their screening is reasonable in the presence of characteristic symptoms of PV such as aquagenic pruritus or unexplained splenomegaly, even if the complete blood count picture is not suggestive of PV.⁴⁶
• •JAK2V617F studies are concordant between peripheral blood and bone marrow; therefore, patients may be screened with either specimen type, but analyzing both blood and marrow is unnecessary.^47,48
• •JAK2 exon 12 mutations are rare in ET or PMF, and their occurrence in PV is almost always associated with the absence of JAK2V617F and the presence of a subnormal serum erythropoietin level.^46,49
• •The incidence of MPL mutations in MPN are too low (see above) to warrant their routine use in MPN diagnosis, except for clarification of equivocal morphology in the diagnosis of ET or PMF.⁷
• •JAK2 and MPL mutations do not occur in healthy subjects or in those with non-clonal causes of myeloproliferation.^7,50,51

Taking the above-mentioned factors into consideration, we have outlined the clinical scenarios in which JAK2 or MPL mutations are indicated (Table 2) and also provided a diagnostic algorithm for PV, ET, and PMF (Figure 1). At present, it is not essential to use quantitative assays for mutation screening, although we prefer to use cell-based quantitative assays because diagnosis certainty is enhanced in the presence of >1% mutant allele burden.⁵⁰ In addition, a ≥50% JAK2V617F allele burden suggests the presence of homozygous mutations, which are atypical for ET,^40,52 but a lower mutant allele burden does not necessarily exclude the presence of homozygously mutated cells and is, therefore, of limited value in distinguishing ET from PV or PMF. Cell-based quantitative assays are also useful to monitor treatment response, including assessment of minimal residual disease after allogenic stem cell transplantation.^53–56

Table 2.

Clinical Indications for Screening JAK2 and MPL Mutations

Mutation	Screening appropriate	Not indicated
JAK2V617F	Erythrocytosis	For purposes of MPN prognostication
	Thrombocytosis	To differentiate one MPN from another
	Bone marrow fibrosis	Testing both blood and bone marrow
	BCR-ABL1-negative granulocytosis
	Unexplained monocytosis
	Unexplained splenomegaly
	Aquagenic pruritus
	Splanchnic vein thrombosis
	Testing either blood or bone marrow
JAK2 exon 12 mutation	JAK2V617F-negative erythrocytosis and low Epo	Before JAK2V617F screening
	Suspected JAK2V617F-negative post-PV MF	In the presence of JAK2V617F
		For diagnosis of ET or PMF
MPL mutation	Thrombocytosis and morphologically equivocal for ET	For diagnosis of PV Morphologically confirmed ET or PMF
	Marrow fibrosis and morphologically equivocal for PMF	JAK2V617F positive

Epo, serum erythropoietin level; MF, myelofibrosis.

Figure 1.

A contemporary diagnostic algorithm for MPNs, including PV, ET, and PMF. Epo, erythropoietin; V617F, JAK2V617F; BM, bone marrow.

Recent studies in PV have suggested the association of higher mutant allele burden with a higher risk of fibrotic transformation⁵⁷ and in PMF a lower mutant allele burden with inferior survival.^58,59 However, the lack of assay standardization across different laboratories undermines the practical translation of such observations, at present. In general, mutation screening outcome does not appear to be influenced by or whether peripheral blood or bone marrow is used as the source of test samples.⁴⁷ Finally, it is important to note that most commercial assays use ≥1% sensitivity level to minimize false-positive test results; however, positive signals in the 0.01% to 1% range should not be discarded, especially in the context of the appropriate clinical scenario.