Detection of the JAK2^sup V617F^ Mutation in Myeloproliferative Disorders by Melting Curve Analysis Using the LightCycler System

Chronic myeloproliferative disorders (MPDs) represent a heterogeneous group of clonal stem cell diseases characterized by abnormal hematopoietic cell proliferation with relatively normal differentiation and maturation. Depending on the specific lineage affected, MPDs may manifest as polycythemia vera, chronic myelogenous leukemia, essential thrombocythemia, or idiopathic myelofibrosis. Although diagnostic criteria based on clinical and morphologic findings have been defined for each entity, the lack of any one unique diagnostic feature makes MPDs difficult to distinguish from reactive bone marrow conditions. The only recurrent genetic abnormality associated with MPDs is t(9;22), which results in the characteristic BCR/ABL fusion gene of chronic myelogenous leukemia.

Recently, several reports have described a specific mutation (V617F) in the Janus nonreceptor tyrosine kinase 2 (JAK2) gene in a significant proportion of patients with MPDs. The somatic mutation, in which a missense mutation changes the amino acid at position 617 of the JAK2 protein from a valine to a phenylalanine residue, occurs in the transformed hematopoietic precursor and has been identified in 65% to 97% of polycythemia vera specimens, 23% to 57% of essential thrombocythemia specimens, and 35% to 57% of myelofibrosis specimens.1-6 This mutation has also been observed in several related leukemic disorders, including 33% of chronic neutrophilic leukemia, 2% of chronic eosinophilic leukemia, 20% of chronic myelomonocytic leukemia, and rare cases of myelodysplastic syndromes and megakaryocytic leukemia.7 Acquisition of JAK2^sup V617F^ has never been identified in healthy controls or patients known to have unrelated bone marrow disorders.8-11 These studies strongly suggest that the presence of JAK2^sup V617F^ may have diagnostic value for MPDs, particularly when the clinical or morphologic findings are equivocal. Furthermore, measuring the proportion of granulocytic cells carrying the JAK2^sup V617F^ mutation may serve as a marker of disease progression or treatment response.

To date, the majority of published studies evaluating JAK2 in patients with MPD have utilized various DNA sequencing platforms to identify the presence of the mutated allele in purified granulocyte fractions of peripheral blood or bone marrow. Although this is an appropriate approach, real-time polymerase chain reaction (PCR) is a more practical technique for routine use in a clinical setting. The real-time PCR assay presented here can be easily integrated into the operations of a molecular diagnostics laboratory. Specific primers and hybridization probes were designed to distinguish the wild-type and mutant JAK2 alleles by LightCycler (Roche Applied Science, Indianapolis, Ind) melting curve analysis. This assay was shown to be a rapid and reliable method for identifying and semiquantitatively measuring the presence of JAK2^sup V617F^ from control cell lines, unfractionated peripheral blood or bone marrow specimens, and archived diagnostic materials.

MATERIALS AND METHODS

Patient Samples

Pathology and hematology records at The Methodist Hospital (Houston, Tex) were retrospectively reviewed to identify patients with a diagnosis of primary MPD, myelodysplastic syndrome, acute leukemia, or reactive hematologic disorder. Test specimens were derived from residual fresh diagnostic material submitted to the molecular diagnostics laboratory or from archived diagnostic material retrieved from the pathology storage facility. Twenty 24- to 48-hour-old peripheral blood specimens were randomly selected from the factor V Leiden and prothrombin G20210A test queue; it was assumed that the clinical suspicion of a hypercoagulable state placed this patient population at an increased risk for having an undiagnosed MPD. No fresh peripheral blood specimens were available for patients with known MPD. Twenty unstained peripheral blood or bone marrow aspirate smears, 20 Wright-stained peripheral blood or bone marrow aspirate smears, and 20 unstained formalin-fixed, paraffin-embedded clot sections were obtained from the pathology archives; the age of these slides ranged from 1 month to 7 years. The 80 specimens were derived from 57 different patients. Homozygous mutant (JAK2^sup V617F^/JAK2^sup V617F^) human erythroleukemia (HEL), and homozygous wild-type multiple myeloma (RPMI8226) cell lines were used as positive and negative controls, respectively. This study was approved by the Institutional Review Board of the Methodist Hospital.

DNA Extraction

The DNA was extracted from the cell lines and patient samples using standard methods. The HEL or RPMI8226 cells were obtained from log-phase cultures; hematopoietic cells were scraped from glass slides or 100

PCR Primers and Hybridization Probes

The PCR primers were designed to flank the guanine to thymine transversion in codon 617 of the JAK2 gene, including forward primer JAKLCFP 5'-AAg CAg CAA gTA TgA TgA gCA A-3' and reverse primer JAKLCRP 5'-AgC TgT gAT CCT gAA ACT gAA-3' (Figure, A). Fluorescence resonance energy transfer probes were designed so the 5' probe overlapped the mutated codon and the 3' probe annealed immediately downstream, including LCRD 5'-640-CAg A*C*A CAT ACT CCA TAA TTT-3' and LCFN 5'-gTA gTT TTA CTT ACT CTC gTC TC-FITC-3', where the asterisk denotes the nucleotide complementary to the wild-type/mutation site (Figure, A).

Real-Time PCR and Melting Curve Analysis

Codon 12 of the JAK2 gene was amplified using the LightCycler platform (Roche Applied Science). Real-time PCR was performed on each specimen using either 5.0 L of purified DNA extract in a total reaction volume of 20 µL that included 4 µL of FastStart DNA Master^sup PLUS^ SYBR Green I 5× reaction master mix (Roche Applied Science), 2.0 µL JAK2LCFP (final concentration 0.5µM), 2.0 µL JAK2LCRP (final concentration 0.5µM), 1.0 µL LCFN (final concentration 0.5µM), 1.0 µL LCRD (final concentration 0.5µM), and 5.0 µL nuclease-free water. The PCR cycle parameters were one initial denaturing step of 95°C for 10 minutes and 55 cycles consisting of 95°C for 10 seconds, 60°C for 60 seconds, and 75°C for 10 seconds. The DNA melting curve analysis was performed by denaturing at 95°C for 10 seconds, annealing at 29°C for 60 seconds, and melting by a transition rate of 0.20°C/ s to 70°C. Melting curves were visually analyzed, and the melting temperature (T^sub m^) of each sample was electronically recorded.

Agarose Gel Analysis and DNA Sequencing

Amplification by real-time PCR was verified by visualization of an appropriately sized 163 base pair band when 5 µL of the real-time PCR product (Roche Applied Science) was applied to a 2% agarose gel and subjected to electrophoresis at 125 m V for 10 minutes. The JAK2 mutation status determined by the realtime PCR assay was initially verified by sequencing. For sequencing, real-time PCR amplicons were generated in the absence of hybridization probes and purified using a QIAquick PCR purification kit (QIAGEN Inc) according to the manufacturer's instructions. Sequencing was performed using the DTCS Quick Start Kit (Beckman Coulter, Inc, Fullerton, Calif) according to manufacturer's instructions with 1 µL of each purified amplicon added to 1.6µM of JAK2LCFP forward primer or JAK2LCRP reverse primer in a Peltier Thermal Cycler 200 (MJ Research Inc/ Bio-Rad Laboratories Inc, Hercules, Calif) through 30 cycles consisting of a denaturation step at 96°C for 20 seconds, an annealing step at 50°C for 20 seconds, and an extension step at 60°C for 120 seconds. The resulting product was resolved by capillary electrophoresis on the CEQ8000 Genetic Analysis System (Beckman Coulter Inc) according to the manufacturer's instructions, and the electropherogram was visually inspected to identify the presence of the wild-type or mutant genotype at codon 617.

RESULTS

Optimization of PCR Amplification and DNA Melting Curve Analysis

The JAK2 LightCycler assay was optimized for real-time PCR amplification and melting curve analysis. Using a standard reaction mix, purified DNA extract from either the HEL cell line, the RPM18226 multiple myeloma cell line, or a 1:1 mixture of each was amplified for 30, 40, 55, and 70 cycles. The ability to visualize the T^sub m^ inflection point of melted DNA was enhanced by plotting the negative first derivative (-dF/dT) versus temperature (T). The -dF/dT peak height ratio was optimal at 55 cycles, and additional cycles did not significantly improve this ratio (data not shown). Similarly, the optimal starting temperature for melting curve analysis was determined by varying the starting temperature from 25°C to 36°C. A 29°C starting temperature was shown to generate optimal results (data not shown). When DNA from the RPMI8226 and HEL cell lines were mixed, the two melting curves observed at approximately 54°C (wild type) and 45°C (JAK2^sup V617F^) could be easily resolved from one another (Figure, B through E). Similarly, comparison of the melting curves generated from amplification of pure HEL or RPM18226 DNA resulted in distinct single curves (Figure, B through E).

Precision and Reproducibility of LightCycler System Melting Curve Analysis

The precision of the JAK2 LightCycler assay was determined by repeating the real-time PCR amplification and DNA melting curve analysis 10 separate times using HEL and RPMI8226 DNA. The JAK2 wild-type amplicon displayed a mean T^sub m^ of 54.28°C and a coefficient of variation of 0.42% (data not shown). The JAK2^sup V617F^ mutant amplicon had a mean T^sub m^ of 45.48°C and a coefficient of variation of 0.44% (data not shown).

The reproducibility of the JAK2 LightCycler assay was determined by repeated melting curve analyses of the same HEL or RPMI8226 DNA sample. Each reproducibly yielded identical -dF/dT versus T curves. The mean T^sub m^ and SD of the wild-type JAK2 amplicon was 54.32°C and 0.12, respectively (data not shown), and the mean and SD of the JAK2^sup V617F^ mutant amplicon was 45.85°C and 0.09, respectively (data not shown).

Sensitivity to Detect the JAK2^sup V617F^ Mutation

The ability of the JAK2 LightCycler assay to identify low concentrations of the V617F allele was evaluated by titration of the homozygous mutant HEL cell line DNA with increasing amounts of the homozygous wild-type RPM18226 cell line DNA. Serial mixtures of HEL and RPM18226 in the ratios of 1:1, 1:2, 1:5, 1:10, 1:20, and 1:40 were amplified by real-time PCR and analyzed by melting curve analysis. In 5 replicate experiments, the 2 respective melting curves were easily distinguishable from the 1:1 to 1:20 mixtures; however, the mutant allele curve was inconsistently detected at the 1:40 dilution (2/5 times) (Figure, B). Thus, the lower limit of detection for JAK2^sup V617F^ was defined as one copy per 20 alleles. As the mutant to wildtype dilution factor increased, the relative area under the mutant curve was observed to be inversely proportional to the area under the wild-type curve (Figure, B).

Detection of the JAK2^sup V617F^ Mutation in Patients With Myeloproliferative Disorders or Reactive Conditions

The JAK2 LightCycler assay was further evaluated with 20 fresh peripheral blood specimens, 20 unstained bone marrow aspirate smears, 20 Wright-stained peripheral blood or bone marrow aspirate smears, and 20 paraffinembedded, formalin-fixed bone marrow clot sections from 57 patients with a previous diagnosis of MPD or a reactive condition. The wild-type allele was identified in each fresh peripheral blood specimen (mean T^sub m^, 54.30°C; SD, 0.20; Table 1, cases 1.01-1.20), unstained smear (mean T^sub m^ 54.51°C; SD, 0.55; Table 1, cases 2.01-2.20), and Wrightstained smear (mean T^sub m^ 54.16°C; SD, 0.53; Table 1, cases 3.01-3.20). The wild-type allele was identified in only half of the processed clot sections (mean T^sub m^ 54.50°C; SD, 0.37; Table 1, cases 4.01-4.20). The JAK2^sup V617F^ mutant allele was identified in none of the peripheral blood specimens (Table 1), 4 of the unstained smears (mean T^sub m^ 45.43°C; SD, 0.04; Figure, C; Table 1, cases 2.06, 2.12, 2.15, and 2.20), 5 of the stained smears (mean T^sub m^ 45.21°C; SD, 0.13; Figure, D; Table 1, cases 3.08-3.09 and 3.15-3.17), and 9 of the clot sections (mean T^sub m^ 45.37°C; SD, 0.15; Table 1, cases 4.01-4.03, 4.10-4.11, 4.13-4.15, and 4.20).

The 18 cases positive for JAK2^sup V617F^ represented patients with essential thrombocythemia (2/3 specimens, 1/2 patients), polycythemia vera (1/1 specimen, 1/1 patient), idiopathic myelofibrosis (9/10 specimens, 6/7 patients), and leukemia transformed from a preexisting MPD, (3/3 specimens, 2/2 patients). These mutant allele frequencies are similar to those of previous reports.1-26 Of note, concordant presence or absence of the JAK2^sup V617F^ mutant allele was observed in all patients having multiple specimens and/or specimen types available for study (Table 2).

COMMENT

Despite their diverse clinical presentation, polycythemia vera, essential thrombocythemia, and idiopathic myelofibrosis traditionally have been classified in a common group termed the myeclopraliferative disorders.13 This system has been the subject of much ongoing debate,14 and although precise diagnostic criteria have been defined for each entity, the lack of a characteristic morphologic or cytogenetic feature makes their distinction from reactive conditions difficult.15,16 The recent identification of the V617F mutation in the Janus kinase 2 gene (JAK2) in a significant proportion of patients with MPD represents the first discriminatory laboratory finding. The absence of this mutation in healthy controls or patients with unrelated bone marrow conditions further supports its use as a diagnostic tool. The majority of previous JAK2 studies have utilized various DNA sequencing techniques to identify the mutant allele.1-5,17 Although appropriate, DNA sequencing is not well suited for routine use in a clinical laboratory; it is cumbersome, time-consuming, and technically demanding. Also, its relatively low sensitivity necessitates the presorting of granulocyte fractions to eliminate nonneoplastic lymphocytes from the DNA pool.1 In comparison, PCR-based methods such as real-time PCR, amplification refractory mutation system PCR, sequencespecific primer single molecule florescence detection, PCR-single-strand conformational polymorphism, and allele-specific PCR represent preferred approaches.18-21 The instrumentation and technical expertise most readily available to typical molecular diagnostics laboratories is real-time PCR.22 Although McClure et al21 recently published a similar melting curve assay, our assay demonstrated the unique ability to retrospectively evaluate archived material and semiquantitatively estimate mutant allele proportion (Figure; Tables 1 and 2). Additionally, this assay could be completed within 3 hours of specimen receipt, allowing for a reasonable turnaround time in the molecular laboratory.

Our LightCycler assay reproducibly detected 1 JAK2^sup V617F^ mutant allele in 20 total alleles. This 5% detection sensitivity is significantly greater than conventional sequencing methods1'1-1 and comparable to pyrosequencing, sequence-specific primer single molecule florescence detection, and amplification refractory mutation system PCR.1,17,19 Thus, it represents an ideal clinical tool for screening the JAK2^sup V617F^ mutation in patients who present with features suggestive of, but not diagnostic for, MPDs. Although a negative test result will not completely exclude the possibility of an MPD, it will indicate that the patient, at most, has a small number of mutant alleles. If clinical suspicion for MPD remains high, these patients should be followed by this assay to examine whether the amount of mutant allele increases over time. Other studies such as complete blood count or marrow evaluation can help to establish or exclude the diagnosis. Allele-specific-PCR assays theoretically may be capable of achieving an even greater analytical sensitivity; however, they are unlikely to result in a significantly better mutant allele detection rate.21

Another advantage of our melting curve assay is its ability to perform retrospective studies on archived specimens. Comparison of the semiquantitative mutant allele estimate to the medical record may show a correlation between JAK2^sup V617F^ proportion and disease progression or treatment response. In the current study, multiple specimens from various time points were available for one patient who had a long history of polycythemia vera that progressed through a postpolycythemic myelofibrosis phase and eventually transformed to acute myelogenous leukemia (Table 2). Comparison of the melting curves demonstrated that this patient had an increasing JAK2^sup V617F^ to wild-type proportion during disease progression, whereas the mutant allele was no longer detectable following allogeneic bone marrow transplant (Figure, E; Table 2). If this preliminary finding is confirmed in a larger patient population, then disease progression may be identified at the molecular level prior to its clinical manifestation and lead to potentially earlier therapeutic interventions. Similarly, a lack of JAK2^sup V617F^ allele detection following treatment may confirm the efficacy of conventional chemotherapeutics or rationally designed small molecule tyrosine kinase inhibitors.

The only limitation of this melting curve assay is its inability to discriminate between clones that are heterozygous or homozygous for the JAK2^sup V617F^ mutation; however, molecular studies have yet to definitively demonstrate that the homozygous state confers a distinct clonal advantage or disease phenotype in MPDs. In the absence of a purified clonal population being available for evaluation, few of the previously described assays are able to unequivocally distinguish a truly heterozygous cell from a homozygous mutant cell in a background of wild-type cells; amplification refractory mutation system PCR and pyrosequencing can predict homozygosity versus heterozygosity, but they cannot perform semiquantitative allele measurements.19,24 Similarly, our LightCycler assay cannot distinguish a mixed population of multiple clones from a pure clonal lineage. Both conditions have been observed in patients with MPD, but their significance remains uncertain.5 If deemed necessary, the zygosity or clonality of a particular specimen could be easily elucidated with the addition of a hematopoietic colony isolation step that selects individual precursor cells for evaluation.25 Although not technically difficult, this would be costly and timeconsuming for a molecular diagnostics laboratory. Further studies are necessary to determine whether this step could provide clinically useful data.

This real-time PCR melting curve assay functioned exceptionally well with fresh or archived peripheral blood and bone marrow aspirate specimens, but its less-impressive performance with clot sections was probably multifactorial. For the 10 cases that did not generate a wild-type or mutant allele curve, review of the clot section demonstrated either significant red blood cell contamination or scant hematopoietic material. Thus, there may have been an insufficient number of nucleated precursor cells present to provide an adequate genomic sample. Also, compared with fresh specimens and archived smears, the additional complexity introduced by formalin fixation and paraffin embedding may have resulted in inefficient DNA extraction or incomplete removal of PCR inhibitors. Additional studies will be necessary to optimize the DNA extraction procedure for clot sections.

In summary, the currently described real-time PCR assay with melting curve analysis represents a suitable molecular diagnostic method for detecting the JAK2^sup V617F^ mutation in clinical specimens. It has a simple sample-processing step, rapid turnaround time, good diagnostic sensitivity, and excellent precision and reproducibility. Furthermore, its ability to use archived materials allows for the retrospective study of JAK2^sup V617F^ in MPDs. These investigations may have significance in the evaluation of disease progression and prognosis.

© 2006 College of American Pathologists Provided by ProQuest LLC. All Rights Reserved.