Single Nucleotide Polymorphism (SNP)-Based Loss of Heterozygosity (LOH) Testing by Real Time PCR in Patients Suspect of Myeloproliferative Disease

During tumor development, loss of heterozygosity (LOH) often occurs. When LOH is preceded by an oncogene activating mutation, the mutant allele may be further potentiated if the wild-type allele is lost or inactivated. In myeloproliferative neoplasms (MPN) somatic acquisition of JAK2V617F may be followed by LOH resulting in loss of the wild type allele. The occurrence of LOH in MPN and other proliferative diseases may lead to a further potentiating the mutant allele and thereby increasing morbidity. A real time PCR based SNP profiling assay was developed and validated for LOH detection of the JAK2 region (JAK2LOH). Blood of a cohort of 12 JAK2V617F-positive patients (n = 6 25–50% and n = 6>50% JAK2V617F) and a cohort of 81 patients suspected of MPN was stored with EDTA and subsequently used for validation. To generate germ-line profiles, non-neoplastic formalin-fixed paraffin-embedded tissue from each patient was analyzed. Results of the SNP assay were compared to those of an established Short Tandem Repeat (STR) assay. Both assays revealed JAK2LOH in 1/6 patients with 25–50% JAK2V617F. In patients with >50% JAK2V617F, JAK2LOH was detected in 6/6 by the SNP assay and 5/6 patients by the STR assay. Of the 81 patients suspected of MPN, 18 patients carried JAK2V617F. Both the SNP and STR assay demonstrated the occurrence of JAK2LOH in 5 of them. In the 63 JAK2V617F-negative patients, no JAK2LOH was observed by SNP and STR analyses. The presented SNP assay reliably detects JAK2LOH and is a fast and easy to perform alternative for STR analyses. We therefore anticipate the SNP approach as a proof of principle for the development of LOH SNP-assays for other clinically relevant LOH loci.


Introduction
Neoplastic disorders may have a variety of underlying molecular and cytogenetic defects. In 2005, the involvement of the JAK2V617F somatic point mutation in myeloproliferative neoplasms (MPN) became apparent. JAK2V617F is caused by a G-to-T transversion in the Janus Kinase 2 (JAK2) gene resulting in a valine-to-phenylalanine (V-F) amino acid substitution at codon 617 [1,2,3,4]. JAK2 is a non-receptor tyrosine kinase involved in the JAK-STAT signalling pathway [5]. The JAK2V617F mutation apparently alters the activity of the autoinhibitory pseudokinase JH2 domain resulting in a constitutive activation of the JAK-STAT signalling pathway, leading to a growth factor independent cell proliferation [6]. The diagnostic value of JAK2 mutational analysis in MPN is now well established and has been endorsed in the classification of haematological malignancies by the World Health Organisation [7].
In addition to neoplasm-promoting mutations, chromosomal imbalance may lead to loss of heterozygosity (LOH) resulting in the deletion of tumor suppressor genes. Besides this, LOH may lead to acquired partial uniparental disomy, excluding the wild-type allele of a cancer-critical gene by which the effect of the mutant oncogenic allele is further potentiated. The latter is commonly found in combination with the JAK2V617F mutation due to LOH of (part of) chromosome 9p [8]. A proposed mechanism which is suggested to give rise to (partial) LOH of chromosome 9p, and thereby JAK2V617F homozygosity, is the mitotic recombination between homologous regions of chromosome 9p in a heterozygous cell [3]. It is further speculated that the daughter cells homozygous for the JAK2V617F mutation gain additional proliferative advantage and outcompete the JAK2V617F heterozygous cells and homozygous wild-type JAK2 cells, thereby leading to an expansion of the JAK2V617F homozygous subclone [3].
Polycythemia Vera (PV) patients with JAK2V617F often carry a homozygous subclone in contrast to Essential Thrombocythemia (ET) in which homozygosity is rare. These homozygous subclones may predominate over time. It has been proposed that it may be more common for PV patients to have a mutant allele burden of .50% than for ET patients [9]. This was confirmed in a large multicentre study during which the JAK2V617F mutational status in patients with PV (n = 323) and ET (n = 639) was analysed.
Of the PV patients, 67.8% was found to be JAK2V617F heterozygous and 32.2% JAK2V617F homozygous. Only 2.2% of ET patients were found to be JAK2V617F homozygous, whereas 40.2% was JAK2V617F-negative and 57.6% JAK2V617F heterozygous [10]. Although the clinical relevance of JAK2V617F allele burden is not yet fully understood; a correlation between disease phenotype and the proportion of mutant alleles has been postulated [11]. Splenomegaly, vascular events and pruritus in PV and splenomegaly, arterial thrombosis and microvessel disease in ET have been clinically correlated to mutant allele burdens [12,13]. In addition to the percentage of mutant allele burden by itself, JAK2V617F homozygosity (through JAK2LOH) is thought to give rise to a more symptomatic myeloproliferative disease (MPD) resulting in several clinical manifestations such as lower platelet counts (PV), more frequent evolution into secondary myelofibrosis, higher risks of cardiovascular events, splenomegaly and higher leukocyte counts [10,14,15,16,17]. Besides this, JAK2V617F homozygosity more often necessitates the use of cytoreductive therapies [10,14]. With regard to ET, the large multicenter study showed also that ET patients harboring homozygous JAK2V617F mutation had a 3.9-fold higher risk of major cardiovascular events as compared to both their heterozygous and wild type counterparts [10].
Quantitative real time PCR is an excellent tool for the quantification of JAK2V617F mutant allele burden and is therefore of clinical importance [18]. Also, determination of the LOH status of (part of) the JAK2 region (JAK2LOH) of chromosome 9p may be clinically relevant due to the possibility of pinpointing those patients carrying a JAK2V617F homozygous subclone, which is thought to give rise to a more symptomatic MPN. Several techniques to determine LOH status have been described, such as typing of STRs/microsatellite DNA analysis [19,20], SNP-based pyrosequencing [19] and Affymetrix SNP microarray [8].
Although these methods provide valuable information on the LOH status, they are often not accessible for routine molecular laboratories -due to the necessity of expensive (pyro)sequencing or array equipment-and often need post-PCR processing increasing contamination risk [8,19,20]. Therefore we developed a JAK2-LOH test for routine use based on real time amplification of SNPs located in and adjacent to JAK2.

Ethics Statement
The presented study was performed retrospectively using archival patient samples which were rendered anonymous before use. The Board of Directors of the Jeroen Bosch Hospital and the JBZ Scientific Advisory Board approved this study. Additionally, the JBZ Scientific Advisory Board waived the need for informed consent from the participants due to the anonymous and retrospective nature of the study. No informed consent from the involved patients was therefore obtained.

JAK2V617F-positive Patient Cohort
To test our SNP assay for detection of JAK2LOH, 12 patients were selected based on JAK2V617F-positive mutational status. This status was determined with a quantitative JAK2V617F real time PCR based method using a PNA oligonucleotide for blocking the wild-type allele [18]. Patients 1 to 6 carried a JAK2V617F mutant allele burden of 25-50%, had a mean age of 7368 years (mean 61 standard deviation) and a male:female ratio of 2:1. Patients 7 to 12 with a mutant allele burden of .50%, rendering JAK2LOH very likely, had an age of 7765 years (mean 61 SD) and a male:female ratio of 1:1. EDTA blood samples used for JAK2V617F mutational status were used for LOH analysis. A nonneoplasmic archival formalin-fixed paraffin-embedded tissue (FFPE-tissue) from each patient was included to establish the germ-line profile.
Although not fully correct due to the heterogenous nature of blood tissue, results were stated as loss of heterozygosity or retention of heterozygosity (ROH) throughout this manuscript.

Patients Suspect of MPN Cohort
Eighty-one patients, suspected for myeloproliferative disease, were included in this study. Inclusion was based on EDTA blood samples collected between December 2007 -December 2009 for JAK2V617F analyses in combination with the availability of archived non-neoplasm related FFPE-tissue. Mean age was 60614 years (mean 61 SD) with a 1:1.3 male to female ratio. In addition to these EDTA-blood samples used for LOH analysis, the same cohort of patients was used to analyze non-neoplasm related FFPE-samples from a wide variety of tissues. These samples were originally taken for routine histological diagnostics and used in this study to determine germ-line SNP and STR profiles.

Tissue Processing
Tissues for histological evaluation were fixed according to a routine procedure in 0.01 mol/L buffered (0.005 mol/L disodium hydrogen phosphate anhydrous and 0.005 mol/L sodium dihydrogen phosphate dihydrate, pH 7.0) 10% formalin, and processed for paraffin embedding using a Tissue-Tek VIP 5 (Sakura, Torrance, USA). The dehydration program consisted of 14 steps of 1 hour under continuous agitation, pressure, vacuum, and heating. At 40uC, two 10% formalin treatments were followed by one 70% (v/v) ethanol treatment, two 96% ethanol treatments, three 100% ethanol treatments, and two 100% xylene treatments. Paraffin embedding was done at 60uC in four 100% paraffin treatments.

Genomic DNA Extraction from FFPE-tissues
Paraffin-embedded tissues were trimmed of paraffin excess and cut into 3 mm-thick sections. Approximately 1 to 1.5 cm 2 of sectioned tissue (a single section or short ribbons depending on the surface per section) was put in 250 mL of digestion solution (digestion solution with proteinase K was prepared by adding 100 mL of proteinase K solution (20 mg/mL; Roche Diagnostics GmbH, Mannheim, Germany) and 10 mL of Tween 20 (Merck BV, Amsterdam, The Netherlands) to 2 mL of TE buffer (1 mmol/L EDTA, and 10 mmol/L Tris-HCl buffer, pH 8.0)) and incubated overnight at 45uC. Proteinase K was inactivated the next day by incubation at 100uC for 15-30 minutes. Afterwards, samples were centrifuged for 2 minutes at 14,000 rcf/G. Genomic DNA was extracted from FFPE-tissues using the EasyMAG NucliSens extraction system (BioMérieux Benelux BV, Zaltbommel, The Netherlands). Two hundred mL digested sample, located beneath a paraffin cap, was added to 2 mL NucliSens lysis buffer, homogenized and incubated for 10 min. at room temperature. The mixture was then added to the EasyMAG vessel and 100 mL of diluted magnetic silica (50 mL silica +50 mL ultrapure water) was subsequently added. The DNA was extracted on the EasyMAG machine (BioMérieux Benelux BV, Zaltbommel, The Netherlands) using the ''Generic 2.0.1'' program. Elution was performed in 200 mL NucliSens Extraction buffer 3.

Genomic DNA Extraction from EDTA-blood
Genomic DNA was extracted from blood using the DNA blood mini kit (Qiagen, Hilden, Germany). Two hundred mL EDTA-blood was added to 200 mL AL buffer, homogenized and incubated for 10 min. at room temperature. Two hundred mL of 96% ethanol (Merck KGaA, Darmstadt, Germany) was added. The mixture was transferred to a QIAamp column and centrifuged for 1 min. at 8,000 rcf/G. The column was put in a new collection tube, 500 mL AW1 buffer was added and centrifuged for 1 min. at 8,000 rcf/G. This procedure was repeated with 500 mL AW2 buffer and the column was centrifuged for 1 min. at 14,000 rcf/G. To remove all ethanol from the column it was put in a new collection tube and then subjected to a dry spin for 1 min. at 14,000 rcf/G. Elution was performed by adding 200 mL EL buffer, incubating for 5 min. at room temperature followed by centrifugation for 1 min. at 8,000 rcf/G.

LOH Analysis by STR Detection
Microsatellite markers D9S1810 and D9S1681 spanned a 0.5 Mbp region of chromosome 9p, including the JAK2 gene ( figure 1A) [1]. This 0.5 Mbp region -Mbp 4.8 to 5.3-is located within the minimal LOH region and downstream of the mitotic recombination breakpoint as described by Kralovics and coworkers [3].
The regions containing D9S1810 and D9S1681 were amplified. The underlined sequences indicate the pig tail that was used to promote adenylated products [3]) and 50 ng of isolated DNA.
DNA isolated from both EDTA-blood (tumor) and FFPE-tissue (germline) from each patient were amplified in the same run using a Veriti FAST Thermal Cycler (Applied Biosystems, Foster City CA, USA) for 12 min. at 94uC followed by 10 cycles of 15 sec. at 94uC, 15 sec. at 60uC, 30 sec. at 72uC, followed by 25 cycles of 15 sec. at 89uC, 15 sec. at 60uC, 30 sec. at 72uC and a final elongation step of 10 min. at 72uC.
Prior to capillary electrophoresis, PCR products were diluted in ultrapure water to obtain a fluorescence signal between 400 and 8000 Fluorescent Units. Capillary electrophoresis and the Fluorescence Unit measurement of each PCR product was performed on a 3130 Genetic Analyzer (Applied Biosystems, Foster City CA, USA) in Hi-Di formamide together with a LIZ-600 size marker (Applied Biosystems, Foster City CA, USA). Allele ratios of tumor material were compared to allele ratios of germline DNA, using Genemapper Software Version 4.0 (Applied Biosystems, Foster City CA, USA).

LOH Analysis by SNP Detection
Ten predesigned TaqManH SNP Genotyping Assays (Applied Biosystems, Foster City CA, USA) were used according to the manufacturer's instructions. Twenty-five mL of PCR, using the in house made JBZ 46master mix contained, 20 mmol/L Tris-HCl, pH 8.4, 50 mmol/L KCl, 3 mmol/L MgCl 2 (prepared from 106 PCR buffer and 50 mmol/L MgCl 2 solution delivered with Platinum Taq polymerase), 0.75 U of Platinum Taq polymerase (Invitrogen BV, Breda, The Netherlands), 4% glycerol (molecular biology grade; Calbiochem, VWR International BV, Amsterdam, The Netherlands), 200 mmol/L of each dNTP (Invitrogen BV), 0.5 mL of Rox reference dye (Invitrogen BV), 1.25 mL of predeveloped assay reagent from the Assays-on-Demand SNP genotyping products (Applied Biosystems, Foster City CA, USA) containing two primers and two MGB TaqMan probes (59 VIC for allele 1, 59 FAM for allele 2 and a 39 black hole quencher for both alleles), and 11.25 mL of target DNA. Real time PCR was performed in three different ABI Prism 7500 FAST SDS (ABI) machines for 1 minute at 95uC, followed by 45 cycles of 3 seconds at 95uC and 30 seconds at 60u.

Statistical Analysis SNP LOH Detection
To determine and visualize the distribution of delta Rn VIC/ FAM ratios (ratio allele 1/allele 2), box-and-whisker plots were generated for each SNP based of the EDTA-blood samples from the heterozygous patients -based on the germline profile generated using the corresponding FFPE-tissues-from the patient cohort that were negative for the JAK2V617F mutation (63/81). The box extends from the 25 th to 75 th percentile of observations, representing the interquartile range (IQR). The bars/whiskers define the +1.5 IQR and 21.5 IQR and indicate the cut off value for normal ratio distribution. Therefore, ratios plotted within 1.5-36IQR are outliers and were considered equivocal with regard to JAK2LOH status. Ratios plotted $36IQR are extremes and were considered to be representative for JAK2LOH. All ratios with an IQR of ,1.5 follow normal distribution and were interpreted as JAK2 ROH.
The cell counts were firstly judged for fit to the normal distribution by using stem-and-leaf plots and quantile-quantile plots. As our data did not follow a normal distribution, Mann-Whitney U tests were performed for comparison of the cell counts between the JAK2ROH and JAK2LOH group.

STR and SNP Selection
STRs D9S1810 and D9S1681 span the 0.5 Mbp minimal 9pLOH region, incorporating the JAK2 gene ( figure 1A), and have been used before by Kralovics and coworkers to investigate 9pLOH [3].
Ten SNPs were selected in silico using SNP browser software Version 4.0 (ABI). Selection was based on allele frequencies of around 0.5 and location within or near the JAK2 gene covering the 0.5 Mbp 9p region in between D9S1810 and D9S1681 (figure 1A and table 1). For real time PCR based amplification and SNP detection, TaqManH SNP Genotyping Assays from Applied Biosystems were used. During technical validation, one (SNP rs7862852) of the ten SNP assays showed cross reactivity of the FAM MGB probe with allele 1 (VIC) and poor amplification of allele 2 (FAM) in the presence of allele 1. Therefore, allele calling could not be performed and TaqManH SNP Genotyping Assay for rs7862852 was excluded from LOH analyses. The other 9 SNP assays generated reliable allele calls and were therefore suitable for inclusion in the JAK2LOH panel ( figure 1A and table 1).

Patient Cohorts
Two patient cohorts were used in this study: a cohort of 12 JAK2V617F-positive patients and a cohort of 81 patients suspect of MPN.
FAM/VIC ratios were plotted per SNP in a scatterplot. Figure 2A shows the scatterplot of SNP rs3780378 for paired paraffin and blood samples of all patients in both cohorts. The data obtained for SNP rs3780378 are representative for the entire SNP panel. Four clusters -homozygous VIC, homozygous FAM, heterozygous FAM/VIC ABI7500FAST machine 1 and 2 and heterozygous FAM/VIC ABI7500FAST machine 3-were observed. FAM/VIC ratios plotted outside these clusters were considered to represent LOH or to be equivocal as was statistically determined using box-and-whisker plots ( figure 2B). In addition, a FAM/VIC ratio of a blood sample from a heterozygous patient present in a homozygous cluster was considered to represent LOH.

Statistical Analysis SNP LOH Detection
Box-and-whisker plots were generated, based on the VIC/FAM ratios from EDTA-blood samples of all heterozygous patients from the patient cohort suspect of MPN that were negative for the JAK2V617F mutation (63/81), for all SNP assays (data not shown). Data from the JAK2V617F-positive cohort samples that were selected based on their mutant allele burdens of 25-50% and .50% JAK2V617F and therefore more likely to contain JAK2LOH were not included. Inclusion of JAK2V617F-positives would lead to a larger variety in VIC/FAM ratios and therefore a biased IQR. This would influence outlier (equivocal) and extreme (JAK2LOH) calling. VIC/FAM ratios generated with the JAK2V617F-positive patients were called (JAK2ROH, equivocal or JAK2LOH) using the measured IQRs of the JAK2V617Fnegative patients from the patient cohort. One representative box-and-whisker plot generated using SNP rs3780378 is shown in figure 2B. Based on these calculations, patient 15, 39, 44 and 61 were found to be extremes (JAK2LOH) as additionally presented in figure 1B.

JAK2V617F-positive Patient Cohort
The cohort of 12 JAK2V617F-positive patients was composed with samples containing mutant allele burdens of 25-50% (patients 1-6) and burdens above 50% (patients 7-12). The results of STR vs. SNP LOH analyses are summarized in cross table 2. Patients 1, 2 and 4 showed STR and SNP results consistent with ROH of the JAK2 region. Patients 5, 7, 9, 10, 11 and 12 showed JAK2LOH in both assays. In patient 3 discrepant results were found in STR vs. SNP testing (ROH/LOH). The STR assay results were indicative of JAK2ROH, whilst the SNP profile revealed different results among the various SNPs. SNPs rs7857730 and rs3780378 revealed the presence of one of the two germ line alleles, while the 3 other informative SNPs were found to be heterozygous. Overall results for patient 5 were ROH/ROH. However, SNP rs7869592 result was found to be equivocal, while the 4 other SNPs showed ROH and the remaining 4 SNPs were non-informative (table 2). LOH of the JAK2 region was additionally shown in patient 8 using the SNP assay while STR results were non-informative ( figure 1B and  table 3). In patient 4, SNP rs7869592 was heterozygous but was found to be equivocal, STR results were indicative for ROH.

Patients Suspect of MPN Cohort
Eighteen patients harbored the JAK2V617F mutation, whereas 63 patients did not. The JAK2LOH/ROH results for both groups, as generated by SNP and STR analysis, are summarized in cross    25, 59 and 69 were found to be non-informative with the STR assay, while SNP analysis showed JAK2ROH. SNP profiles of patients 32, 55 and 93 were considered non-informative due to an incomplete SNP profile (4, 7 and 5 out of 10 SNPs, respectively) of the reference FFPE-sample (due to low DNA loads) in combination with being homozygous for all interpretable SNPs. These patients showed JAK2ROH in the STR assay (table 3).

Discussion
We have developed a SNP based assay, which is easy to perform and reliably detects JAK2LOH. Ten SNPs located on 9p within the JAK2 region were selected in silico based on incidence around 0.5 in Caucasian (CEU, CEPH (Centre d'Etude du Polymorphisme Humain) from Utah), Chinese (CHB, Chinese from Beijing), Japanese (JPT, Japanese from Tokyo) and African (YRI, Yoruba from Ibadan Nigeria) populations. Technical validation of the accompanying SNP assays revealed that nine SNP assays fulfilled the criteria for inclusion -clear distinction of homo2/ heterozygous clusters, allele frequency in investigated population around 0.5 (data not shown) and sufficient sensitivity-in the JAK2LOH/ROH panel. SNP rs7862852 was excluded from the recommended panel of SNP markers due to poor performance.
A second technical issue with regard to the SNP assay was encountered when three different ABI7500FAST systems were employed. Two distinct groups of homozygotes were observed for machine 1&2 vs. machine 3 (figure 2). For each ABI7500FAST system all dyes employed (including FAM and VIC) are calibrated individually once every three months in our diagnostic setting.
Variations introduced during such a calibration are inevitable as well as those caused by differences in light intensity generated by the integrated lamp which is replaced after 2,000 burning hours. These technical issues may induce inter-machine variability. To prevent misinterpretation of JAK2LOH/ROH status, we recommend analyzing EDTA-blood samples and corresponding FFPEtissue on one ABI7500FAST as a control within the same run.
When interpreting the results with regard to JAK2LOH and ROH status, it should be taken into account that blood is a heterogeneous tissue. The demonstrated technique will therefore only detect JAK2LOH in those patients where the homozygous subclone is abundant enough. The assess the range in which homozygous subclones are abundant enough to be detected by the SNP assay, a serial dilution of DNA of a healthy homozygous patient and DNA of a healthy heterozygous patient (for SNP rs7849191) was made and tested. The 5%, 10% and $20% (20, 30, 40 and 50%) dilutions yielded results that were interpreted using the generated cut off values as ''ROH'', equivocal with regard to ''LOH status'' and indicative of ''LOH'', respectively (results not shown). Indicating a sensitivity of at least 20% homozygous cells in a heterozygous background using representative SNP rs7849191.
With regard to the interpretation of the STR results, difficulties were encountered when analyzing the data automatically. When allelic ratios were called by the Genemapper Software Version 4.0 (ABI), LOH/ROH calls were found to be doubtful in cases where DNA loads were aberrant -low loads observed in several FFPEtissues and high loads in some blood samples-. The software did not take into account that signals .8,000 or with very low fluorescent units negatively influenced allelic ratios, resulting in false LOH calls. Therefore, an additional manual analysis of all STR results was performed to determine STR JAK2 status. To assure that manual interpretation was objective, an LOH index was calculated. This index was calculated by determining the intensity of the PCR products, which was used to calculate the intensity ratio between the peaks of both alleles in the tumor sample. Dividing this ratio by the same ratio found in the normal tissue yielded the LOH index. When this LOH index was below 0.75, this was interpreted as LOH, whereas an LOH index of .0.75 was interpreted as ROH.
Considering the STR assay to be the golden standard, the SNP assay's sensitivity and negative predictive value were 100% for both the JAK2V617F-positive patient cohort and the patients suspect of MPN cohort. Positive predictive values of the SNP assay were 85.7% and 100% and the specificity was 80% and 100% in the respective cohorts. Patient 3 showed no signs of JAK2LOH  using the STR assay. Analyses of SNPs close to both STR loci also showed ROH. However, two SNPs in between the two STR loci in relative close proximity to JAK2V617F (rs7857730 and rs3780378 10-15 Kb and 35-50 Kb downstream, resp.) were not indicative of ROH. Rs7857730 and rs3780378 clearly showed JAK2LOH. It is likely that a small deletion, point mutation or a small recombination event around rs7857730 and rs3780378 accounts for these observations. Of the 81 patients in the patient cohort suspect of MPN, JAK2LOH was found in 6 patients (7%), i.e. one-third (6/18) of the JAK2V617F-positive patients. Amongst the JAK2V617Fnegative patients, no cases of JAK2LOH were found.
That JAK2LOH was found in 2/5 patients from the JAK2V617F-positive cohort with a JAK2V617F mutant allele burden below 50% is in line with the findings of Scott and coworkers, describing that a large number of PV patients harbor homozygous subclones, which may predominate over time [9]. Moreover, it indicates that next to JAK2V617F quantification the determination of JAK2LOH status provides additional information and could be of value.
In patient 4, rs7869592 was the only informative SNP and was found to be equivocal. Patient 38-positive for JAK2V617Fyielded an equivocal result for SNP rs7869592, SNP rs10815188 and D9S1681 were representative of JAK2ROH. All other markers were non-informative due to homozygosity. The overall result was therefore concluded to be equivocal for patients 4 and 38 ( figure 1B, table 3). Patient 44 yielded results indicative of JAK2LOH for D9S1810 and 8/9 SNPs, whereas D9S1681 and rs10815188 yielded results indicative of JAK2ROH ( figure 1B). The mechanism behind this observation is unclear. A possible explanation could be an inversion of the D9S1681/rs2224572 locus that might have arisen during the process of recombination. In patient 61, the result from marker D9S1681 was indicative of JAK2LOH, whereas D9S1810 yielded a JAK2ROH result. Results from 6/9 SNPs were indicative of JAK2LOH, the result from one SNP was equivocal (figure 1B). We presume that this aberrant observation is due to mutation of the concerning JAK2 region.
In both patient cohorts the STR assay resulted in a total of 9 non-informative homozygous profiles, of which patient 8 and 68 were found in the SNP assay to harbor cells with JAK2LOH. Analysis of the cohorts using the SNP assay resulted in 3 noninformative homozygous profiles, of which the STR results were representative of JAK2ROH. Thus the SNP assay has a higher success rate than the STR assay in this population. Caution is however advised when handling samples with low DNA loads. Additionally, the SNP assay provides more detailed information with regard to ROH/LOH status of the JAK2 gene and surrounding loci.
In total, in 14 patients the SNP assay results were indicative of JAK2LOH (including patient 3, where a small part of the JAK2 region was affected; figure 1B and table 2) with an overall mean age of 7467 years and a male:female ratio of 1.00:0.75. The 30 JAK2V617F-positive patients consisted of 8 PV patients, 7 ET patients, 1 IMF patient and 14 patients with a non-classified MPN. The occurrence of JAK2LOH amongst the different disease phenotypes was: PV 100% (n = 8), ET 0% (n = 7), IMF 0% (n = 1) and non-classified MPN 42.9% (n = 14). Similar distributions were previously found showing that JAK2LOH is a common event in PV and the contrary, is very rare in ET and IMF [9,10,11,12,13,14,15,16,17]. The fact that within the group of JAK2V617F-positive patients, leukocyte and erythrocyte counts were significantly higher in patients with JAK2LOH in comparison to patients with JAK2ROH may indicate a more symptomatic disease in the first group. This is in line with previous findings where the higher mutant allele burden was found to be correlated to lower platelet counts, higher incidence of splenomegaly and larger spleen size, cytoreductive therapy needed in a greater number of patients and higher white blood cell counts in PV and ET. A higher incidence of thrombotic events is also observed in PV and ET, particularly with mutant allele burdens above 50%. Significantly higher chances of fibrotic transformation have been observed in both PV and ET [10,14,15,16,17]. Additionally, MPN in patients 10 and 44 with JAK2LOH evolved into leukemia; patient 10 suffered from AML and patient 44 suffered from B-CLL in combination with AML and subsequently deceased. No leukemic transformation was observed in the JAK2ROH group.
Because the SNP assay relies on real time PCR, the technique is fast, easy to perform and requires less hands-on time in comparison to STR analysis -1.5 hours vs. 3 hours-. In addition, the SNP assay does not require the use of expensive sequencing equipment, is therefore more easily accessible for smaller routine molecular diagnostic laboratories. No post-PCR processing is necessary, diminishing contamination risks. Besides the current application of JAK2LOH testing, the demonstrated SNP technique may prove useful as a diagnostic strategy for determination of LOH status in other neoplastic entities such as 18q LOH in colorectal cancer [21,22,23]. In addition to the in this study described use of archival FFPE-tissues to generate germline SNP profiles, real time PCR based SNP profiling was described to be suitable in combination with buccal swabs [24], providing moreeasy-to-obtain reference material.
In conclusion, we have developed a novel and easily accessible SNP based assay to reliably determine JAK2LOH status.