- Top of page
Polycythaemia vera (PV) is closely associated with both an acquired activating mutation of the JAK2 tyrosine kinase (JAK2V617F) in granulocyte-derived DNA and increased granulocyte polycythaemia rubra vera-1 (PRV-1) expression. In order to explore the correlation between these two biological markers and compare their diagnostic utility, mutation analysis for JAK2V617F and quantitative measurement of granulocyte PRV-1 expression were performed on the same study sample from 100 participants: 38 with PV, 22 with essential thrombocythaemia (ET), 10 with agnogenic myeloid metaplasia (AMM), 19 with secondary polycythaemia (SP) and 11 healthy volunteers. The respective overall (homozygous) JAK2V617F mutational frequencies were 95% (26%), 55% (0%), 30% (0%), 0% and 0%. The corresponding figures for increased PRV-1 expression were 89%, 18%, 20%, 21% and 9%. In patients with either ET or AMM, the likelihood of detecting JAK2V617F was significantly higher in the presence of an increased PRV-1 expression (83% vs. 38%; P = 0·05). Similarly, in patients with PV, homozygous as compared with heterozygous JAK2V617F correlated with higher levels of PRV-1 expression (P = 0·11). The present study suggests an allele dose-dependent effect of JAK2V617F on granulocyte PRV-1 expression. However, compared with the PRV-1 assay, mutation screening for JAK2V617F displayed greater accuracy in distinguishing PV from SP.
The majority (65–97%) of patients with polycythaemia vera (PV) carry an acquired activating mutation of the JAK2 tyrosine kinase (JAK2V617F) in DNA derived from granulocytes (Baxter et al, 2005; James et al, 2005; Jones et al, 2005; Kralovics et al, 2005; Levine et al, 2005; Zhao et al, 2005). The same mutation is also found frequently in other BCR/ABL-negative myeloproliferative disorders (MPD) including essential thrombocythaemia (32–57%) and agnogenic myeloid metaplasia (AMM) (43–50%) (Baxter et al, 2005; James et al, 2005; Jones et al, 2005; Kralovics et al, 2005; Levine et al, 2005) and less frequently in chronic myelomonocytic leukaemia, myelodysplastic syndrome, systemic mastocytosis, chronic neutrophilic leukaemia, eosinophilic disorders and atypical MPD (Jelinek et al, 2005; Jones et al, 2005; Steensma et al, 2005). However, to date, JAK2V617F has not been reported in either normal controls (Baxter et al, 2005; James et al, 2005; Kralovics et al, 2005; Levine et al, 2005) or patients with secondary erythrocytosis (James et al, 2005; Kralovics et al, 2005). Therefore, detection of the JAK2V617F mutant allele might carry a strong positive-predictive value in distinguishing MPD, including PV, from non-clonal conditions, such as secondary polycythaemia (SP).
Previous studies have also demonstrated a close association between PV and increased granulocyte polycythaemia rubra vera-1 (PRV-1) expression that was neither invariable (test sensitivity ranged from 69% to 100%) nor exclusive (a substantial minority of patients with essential thrombocythaemia (ET), AMM and other myeloid disorders also display the specific abnormality) (Temerinac et al, 2000; Kralovics et al, 2003; Liu et al, 2003; Florensa et al, 2004; Koch et al, 2004; Passamonti et al, 2004; Tefferi et al, 2004; Sirhan et al, 2005). In contrast to that observed with mutation analysis for JAK2V617F, increased PRV-1 expression has been reported in both normal controls and patients with SP (Passamonti et al, 2004; Sirhan et al, 2005). On the contrary, both JAK2V617F and increased granulocyte PRV-1 expression display a significantly higher prevalence in PV and postpolycythaemic myeloid metaplasia as opposed to other MPD, suggesting a correlation between the two markers (Tefferi et al, 2004; Baxter et al, 2005; James et al, 2005; Kralovics et al, 2005; Levine et al, 2005). The present study explored this possibility and compared the diagnostic utility of the two laboratory markers in distinguishing PV from SP.
- Top of page
The current study was approved by the Mayo Clinic institutional review board (IRB) and involved 100 consecutive patients in whom peripheral blood samples were collected between April 2003 and July 2004 under an IRB-approved protocol. The diagnoses of PV, ET and AMM were made according to the World Health Organisation (WHO) diagnostic criteria (Jaffe et al, 2001). The designation of SP required the presence of a comorbidity known to be associated with SP as well as a bone marrow examination that was felt to be inconsistent with a primary myeloid disorder. Bone marrow histology in all cases was reviewed by Mayo Clinic haematopathologists and re-reviewed by one of the authors.
Both JAK2 mutational analysis and measurement of PRV-1 expression were performed on the granulocyte fraction of peripheral blood. Double density gradient centrifugation (Histopaque-1077TM layered over Histopaque-1119TM; Sigma Diagnostics, St Louis, MO, USA) was used to separate out the granulocyte cell layer from each sample. Genomic DNA was extracted using QIAamp Blood Mini Kit (Qiagen, Valencia, CA, USA) and amplified by polymerase chain reaction (PCR). Each 50-μl PCR reaction contained approximately 25 ng of DNA template, 5 μl 10X Roche Buffer (final concentration of MgCl2: 1·5 mmol/l−1), 1·5 U Taq polymerase (Roche, Indianapolis, IN, USA), 0·8 mmol/l−1 dNTPs (Roche), and 20 pmol/l−1 each of sense and antisense primers (5′-TGCTGAAAGTAGGAGAAAGTGCAT-3′ and 5′-TCCTACAGTGTTTTCAGTTTCAA-3′ respectively).
The PCR cycling parameters were: one cycle of 94°C for 2 min; 35 cycles of 94°C for 30 s, 52°C for 40 s and 72°C for 40 s; followed by one cycle of 72°C for 2 min. PCR products were cleaned with QIAquick PCR purification Kit (Qiagen). Fluorescent dye chemistry sequencing was performed using the same primers used for amplification, on an ABI Prism 3700 DNA Analyzer (Applied Biosystems, Foster City, CA, USA). Sequencher 4.2 (Gene Codes Corporation, Ann Arbor, MI, USA) and GenBank accession NM_004972 (JAK2 mRNA) and the corresponding region from the NC_000009 chromosome 9 contig were used for sequence analysis.
The PRV-1 mRNA level was quantified by real-time reverse transcription-PCR according to the previously published methods (Klippel et al, 2003). To standardise results, the experiment was run as a relative quantification assay, incorporating the housekeeping glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene transcript. The mean cycle threshold (CT) value for PRV-1 was divided by the mean CT value for GAPDH, creating a PRV-1/GAPDH ratio. Consequently, a low PRV-1/GAPDH CT ratio indicated increased PRV-1 mRNA level. PRV-1 gene expression was considered elevated in the presence of a PRV-1/GAPDH ratio of 1·17 or lower (Klippel et al, 2003).
Statistical comparison between categorical variables was performed by chi-squared statistics. Comparison between categorical and continuous variables was performed by either the Mann–Whitney U-test or Kruskal–Wallis test. All data were analysed by using sas software (SAS, Inc., Cary, NC, USA).P < 0·05 were considered statistically significant.
- Top of page
Among the 100 study patients, 38 had PV, 22 had ET, 10 had AMM, 19 had SP and 11 were healthy volunteers. Comorbidities associated with SP included obstructive sleep apnoea with or without additional lung disease or tobacco use in five patients, treatment with testosterone in four patients, heavy tobacco use with or without associated lung disease in three patients, haemoglobinopathies with increased oxygen affinity in two patients, chronic obstructive pulmonary disease in two patients, and one patient each with atrial septal defect and pulmonary stenosis, renal cyst and high altitude habitat.
Baseline patient characteristics and the results of both JAK2V617F mutation analysis and real-time PCR measurement of granulocyte PRV-1 transcript levels are presented in Table I. The age distributions were similar among the different disease categories but there were more males represented in the AMM and SP disease category. However, currently there is no evidence with regard to gender bias in terms of either the presence of the JAK2V617F mutation in AMM or neutrophil PRV-1 expression in SP. The same can be said regarding disease duration in general although two studies have suggested a longer disease duration associated with homozygous JAK2V617F in MPD (Kralovics et al, 2005; Levine et al, 2005).
Table I. Baseline patient characteristics and summary of JAK2V617F mutation analysis as well as quantitative measurement of granulocyte PRV-1 expression in 100 patients with polycythaemia vera (PV), essential thrombocythaemia (ET), agnogenic myeloid metaplasia (AMM), secondary polycythaemia (SP) or healthy volunteers.
| ||PV (n = 38)||ET (n = 22)||AMM (n = 10)||SP (n = 19)||Controls (n = 11)|
| Range||18–81||18–81||41–67||23–77|| |
| Sex (f/m)||20/18||14/8||1/9||3/16||N/A|
|Disease duration (months)|
| Range||0–290||0–152||0–348||0–120|| |
| % with increased value*||89||18||20||21||9|
| Absent (%)||2 (5)||10 (45)||7 (70)||20 (100)||11 (100)|
| Present (%)||36 (95)||12 (55)||3 (30)||0||0|
| Homozygous (%)||10 (26)||0||0|| || |
As expected, increased neutrophil PRV-1 expression had the highest frequency in PV (89%) and was less prevalent in the other MPD including ET (18%) and AMM (20%). More importantly, abnormal levels were also registered in 21% of patients with SP and one of 11 (9%) healthy volunteers. In contrast, the JAK2V617F mutation was not detected in the absence of clonal myelopoiesis (a test specificity of 100%). In the current study, we found the mutational frequency in PV to be over 90% providing a test sensitivity that was comparable with that of neutrophil PRV-1 expression. However, both biological markers were suboptimal in distinguishing PV from other MPD (Table I).
There was a significant across the board correlation between the presence of JAK2V617F mutation in MPD and neutrophil PRV-1 expression (Fig 1, P = 0·0001). When the analysis was restricted to patients with either ET or AMM, patients with the mutation (n = 15) compared with those without (n = 17) were significantly more likely to display increased PRV-1 expression (P = 0·001). Similarly, the 10 PV patients who were homozygous for the mutant allele had a higher PRV-1 transcript level compared with the 26 PV patients who were heterozygous for the mutation, although the difference did not reach statistical significance (P = 0·11).
Figure 1. Distribution of granulocyte polycythaemia rubra vera-1 (PRV-1) and glyceraldehyde-3-phosphate dehydrogenase ratio in 70 patients with myeloproliferative disorders including 38 with polycythaemia vera, 22 with essential thrombocythemia and 10 with agnogenic myeloid metaplasia. The figure displays a significantly altered PRV-1 expression based on mutation status for JAK2; wild-type versus heterozygous JAK2V617Fversus homozygous JAK2V617F. Boxes enclose values within the first and third quartile ranges divided by a line representing the median. Whiskers indicate 1·5 × the interquartile range above and below the 75th and 25th percentile.
Download figure to PowerPoint
- Top of page
The results of the current study confirm a high positive-predictive value (100%) for MPD in the presence of JAK2V617F and support the incorporation of mutation screening for JAK2V617F during the evaluation of polycythaemia as well as thrombocythaemia (Fig 2 and 3). In this regard, its value might supersede that of the PRV-1 assay and other specialised tests for distinguishing PV from SP including platelet-rich plasma serotonin level (Koch et al, 2004), endogenous in vitro erythroid colony formation (Reid, 1987), and megakaryocyte/platelet Mpl expression (Moliterno et al, 1998; Tefferi et al, 2000), each of which have important limitations. Furthermore, diagnostic accuracy might be enhanced by the use of more sensitive techniques that detect JAK2V617F in minor clonal cell populations, such as allele-specific PCR (Baxter et al, 2005; Jones et al, 2005). However, it is underscored that mutation screening for JAK2V617F is not helpful in distinguishing the different categories of chronic myeloid disorders. For comparative purposes, we have included the current WHO criteria for the diagnosis of PV (Table II) and ET (Table III).
Figure 2. A diagnostic algorithm for polycythaemia vera (PV) that incorporates mutation screening for JAK2V617F. PV-characteristic features include increased leucocyte alkaline phosphatase score, thrombocytosis, leucocytosis, splenomegaly, thrombosis, pruritus and erythromelalgia. MPD, myeloproliferative disorder; CBC, complete blood count. *Please note that the current algorithm has not been validated by a systematic study and is based on a consensus working procedure that is currently in use at the authors’ institutions. Furthermore, the algorithm assumes the use of a properly validated modern assay for serum erythropoietin determination (Ma et al, 1992; Mossuz et al, 2004).
Download figure to PowerPoint
Figure 3. A diagnostic algorithm for primary thrombocythaemia (i.e. clinically not consistent with reactive thrombocytosis). ET, essential thrombocythaemia; MPD, myeloproliferative disorder; FISH, fluorescence in situ hybridisation; RT-PCR, reverse transcription-polymerase chain reaction. *Please note that the current algorithm has not been validated by a systematic study and is based on a consensus working procedure that is currently in use at the authors’ institutions.
Download figure to PowerPoint
Table II. World Health Organisation (WHO) criteria for polycythaemia vera (PV).
|1. Elevated red cell mass >25% above mean normal predicted value, or hemoglobin >18·5 g/dl in men, 16·5 g/dl in women or >99th percentile of method-specific reference range for age, sex, altitude of residence.|
|2. No cause of secondary erythrocytosis, including|
| a. Absence of familial erythrocytosis|
| b. No elevation of erythropoietin because of;|
| i. Hypoxia (arterial pO2 ≤ 92%)|
| ii. High oxygen affinity haemoglobin|
| iii. Truncated erythropoietin receptor|
| iv. Inappropriate erythropoietin production by tumour|
|4. Clonal genetic abnormality other than Philadelphia chromosome or BCR/ABL fusion gene in marrow cells|
|5. Endogenous erythroid colony formation in vitro|
|1. Thrombocytosis >400 × 109/l|
|2. Leucocytosis >12 × 109/l|
|3. Bone marrow biopsy showing panmyelosis with prominent erythroid and megakaryocytic proliferation|
|4. Low serum erythropoietin levels|
Table III. World Health Organisation (WHO) criteria for essential thrombocythaemia (ET), Jaffe et al (2001).
|1. Sustained platelet count ≥600 × 109/l|
|2. Bone marrow biopsy specimen showing proliferation mainly of the megakaryocytic lineage with increased numbers of enlarged, mature megakaryocytes|
|Criteria of exclusion|
|1. No evidence of polycythaemia vera (PV)|
| a. Normal red cell mass or hemoglobin <18·5 g/dl in men, 16·5 g/dl in women|
| b. Stainable iron in marrow, normal serum ferritin or normal mean cell volume (MCV)|
| c. If the former condition is not met, failure of iron trial to increase red cell mass or hemoglobin levels to the PV range|
|2. No evidence of chronic myeloid leukaemia|
| a. No Philadelphia chromosome and no BCR/ABL fusion gene|
|3. No evidence of chronic idiopathic myelofibrosis|
| a. Collagen fibrosis absent|
| b. Reticulin fibrosis minimal or absent|
|4. No evidence of myelodysplastic syndrome|
| a. No del(5q), t(3;3)(q21;q26), inv(3)(q21q26)|
| b. No significant granulocytic dysplasia, few if any micromegakaryocytes|
|5. No evidence that thrombocytosis is reactive because of|
| a. Underlying inflammation or infection|
| b. Underlying neoplasm|
| c. Prior splenectomy|
The present study also demonstrated an allele dose-dependent significant association between the presence of JAK2V617F and increased PRV-1 expression (Fig 1). In patients with either ET or AMM, for example, five of the six patients (83%) with increased PRV-1 expression also carried the JAK2V617F mutant allele as opposed to only 10 of the 26 (38%) with normal PRV-1 expression (P = 0·05). Similarly, PV patients with homozygous JAK2V617F, compared with PV patients having heterozygous mutations, displayed higher levels of granulocyte PRV-1 expression (P = 0·11). Taken together, the particular observation suggests a JAK2-mediated effect on granulocyte PRV-1 expression and is consistent with previously reported observations that have linked JAK2V617F with the PV phenotype including the induction of erythrocytosis in mice after transplanting them with murine bone marrow cells transduced with murine JAK2V617F (James et al, 2005), the demonstration of JAK2V617F-mediated cytokine hypersensitivity in cell lines (James et al, 2005; Kralovics et al, 2005; Levine et al, 2005), the invariable presence of JAK2V617F in epo-independent erythroid colonies (Baxter et al, 2005), and the association in ET of the mutant allele with higher haemoglobin levels at diagnosis (Wolanskyj et al, 2005).
We entertained two possibilities, without the exclusion of others, to explain the significant association between homozygous JAK2V617F and PRV-1 expression. In a recent communication, serial analysis of archived bone marrow cells in JAK2V617F homozygote patients disclosed a gradual change from a heterozygous to homozygous mutational status over time that suggested clonal dominance rather than a two-step molecular event (Tefferi et al, 2005). As such, approximately half of the neutrophils in JAK2V617F‘heterozygotes’ might not be clonally involved and thus do not contribute to the PRV-1 signal. The second explanation recalls previous observations regarding increased PRV-1 expression in several reactive conditions as well as in growth factor-stimulated granulocytosis that raised the possibility of the phenomenon being a marker of neutrophil activation (Passamonti et al, 2004). Accordingly, the JAK2V617F-associated increase in PRV-1 might represent a recapitulation of cytokine-associated neutrophil activation that is known to be mediated through the JAK-signal transducer and activator of transcription signalling pathway (Al-Shami et al, 1998). Finally, two recent communications have also demonstrated the close association between PRV-1 expression and JAK2V617F in MPDs (Goerttler et al, 2005; Jones et al, 2005).