Introduction
Myeloproliferative neoplasms (MPNs) are caused by an increased production and clonal expansion of a subset of progenitor stem cells, the CD34 multipotent hematopoietic stem cells.1 These neoplasms arise owing to somatic pathogenic variants (referred to as mutations for convenience henceforth in this text) that result in myeloid proliferation. MPNs are broadly classified into seven different types: chronic myeloid leukemia (CML) i.e., Philadelphia (Ph)-positive MPNs, polycythemia vera (PV), primary myelofibrosis (PMF), essential thrombocythemia (ET), chronic neutrophilic leukemia, chronic eosinophilic leukemia-NOS and MPN-unclassifiable.2, 1 The classical Ph-negative myeloproliferative neoplasms (MPNs), namely PV, PMF, and ET, harbor driver mutations in the Janus Kinase 2 (JAK2),3, 4, 5 myeloproliferative leukemia (MPL),6 and calreticulin (CALR)7, 8 genes. The detection of mutations in these genes has been incorporated into the recent World Health Organization (WHO) diagnostic criteria for MPNs in conjunction with other clinical, hematological, and morphological findings.9 80-90% of Ph-negative chronic MPNs harbor mutations in these genes. At times, subclassifying different MPNs into various subtypes is difficult owing to morphological overlap in disease presentation. Moreover, it is challenging to differentiate between clonal and reactive proliferations. Therefore, the detection of these somatic mutations aids the clinician in patient management. JAK2 mutations are observed in the majority of cases in the Indian subcontinent;10 mutations in the MPL and CALR genes account for 30-40% of Ph-negative chronic MPNs.6 Targeted capillary (or Sanger) sequencing can be used to check these mutations; however, this method is limited by its relatively low sensitivity (20%) for the detection of MPNs as certain low-depth mutations may be missed.11 Owing to a need for clinicians to analyze all three genes simultaneously, we have developed an in-house next generation sequencing (NGS) panel that can simultaneously detect mutations in the targeted regions of these three genes and also quantify the relative allele frequency (mutation depth). This provides clinicians with the ability to monitor response to drugs or bone marrow transplantation. Here, we describe the diagnostic utility of this assay for the molecular analysis of clinically suspected MPNs.
Materials and Methods
Patients and samples
Clinical diagnosis: Samples of patients clinically suspected of having Ph-negative MPNs were sent for analysis by the referring hematologist. The clinical features, peripheral blood and bone marrow findings were not available to us.
DNA extraction
Peripheral blood (2-3 ml) was extracted from individuals with suspected Ph-negative MPN and stored in K2-ethylenediaminetetraacetic acid (EDTA) vacutainers (Becton Dickinson, MD, USA) at 4°C until extraction. Genomic DNA was extracted from 150 μl of blood and eluted in 200 μl of elution buffer using the DNeasy Blood and Tissue Kit (Qiagen, Germany) as per the manufacturer’s instructions.
NGS library preparation
NGS library preparation was performed as described in Ramanan et al.12 Briefly, NGS primers for targeted regions of JAK2 exons 12 and 14, CALR exon 9, and MPL exon 10 were designed using Primer Express v2.0 (Applied Biosystems, CA, USA). Custom “high purity salt-free” primers that contained the target sequences along with NGS adapter tails were synthesized from Integrated DNA Technologies (IDT, IA, USA). A multiplex PCR containing four primer pairs (one primer pair for each: JAK2 exons 12 and 14, CALR exon 9, and MPL exon 10) was performed followed by purification and an indexing PCR step. The pooled purified library at a final concentration of 15 pM with 5% phiX (Illumina, CA, USA), was loaded onto an Illumina MiSeq v2 cartridge (Illumina, CA, USA) and run on an Illumina MiSeq next generation sequencer (Illumina, CA, USA) in 2*250 mode. The targeted depth for each region was approximately 5000 reads.
Bioinformatics
NGS data were analysed using an in-house developed bioinformatics pipeline. The sequences were aligned to the reference GRCH37/hg19 genome and the resulting BAM files were analyzed for variants using the GATK Variant Caller. The BAM and VCF files were visualized using GenomeBrowse v2.1.1 (Golden Helix, MT, USA).
Results
Demographics
In this study, we tested the samples of 274 patients with clinical symptoms indicative of MPNs. Of the 274 patients tested, 190 (69.3%) were male and 84 (30.7%) were female. Eight (2.9%) patients were aged 18 years or younger and 266 (97.1%) patients were above 18 years; 129 (47%) of the 274 patients tested were aged above 50 years.
Prevalence of JAK2, CALR, and MPL mutations
A detailed distribution of the different mutations detected is shown in Figure 1. The average depth for each region was approximately 3000 reads. The JAK2 V617F mutation was detected in 17.5% (n = 49) of the suspected 274 samples (representative snapshot shown in Figure 2). The rare JAK2 exon 12 mutation c.164_1616 delins ATT (H5338Q, K539L) was detected in one patient (0.4%) (Figure 3). 49% of the samples harboring mutations in the JAK2 gene exhibited mutation depths greater than 50% (Table 1). Among the 274 patients, CALR mutations were detected in 31 (11.3%) patients. Type 1 CALR mutations 7, 13 were observed in 17 patients and type 2 CALR mutations were observed in 10 patients (representative images in Figure 4, Figure 5 respectively). In addition, we observed 12-bp, 31-bp, and 46-bp deletions in exon 10 of the CALR gene in one sample each (0.4% each). Of note, 58% of the samples harboring mutations in the CALR gene exhibited mutation depths below 20% and only one sample exhibited mutation depth above 50%. Two samples exhibited MPL exon 10 mutations (W515L and W515R). Moreover, both mutations were observed at mutation depths below 10%. One patient exhibited co-occurrence of mutations in CALR exon 9 and MPL exon 10.14 The CALR mutation was a type 1 mutation observed at a mutation depth of 9%. More than 50% of the positive samples belonged to patients 50 years and above (Table 2). Patients aged above 50 years constituted 87.75% and 64.51% of all JAK2- and CALR-positive samples respectively.
Table 1
Mutation depths observed in the JAK2, MPL, and CALR genes
|
Mutation Depth |
Up to 5% |
5 to 20% |
20 to 50% |
Above 50% |
|
JAK2 |
4.1% (n = 2) |
8.2% (n = 4) |
38.8% (n = 19) |
49% (n = 24) |
|
CALR |
3.2% (n = 1) |
54.8% (n = 17) |
38.7% (n = 12) |
3.2% (n = 1) |
|
MPL |
- |
100% (n = 2) |
- |
- |
Table 2
MPN positivity by age
Discussion
BCR-ABL1-negative MPNs are characterized by constitutive activation of the JAK-STAT signaling pathway. This is owing to the presence of driver mutations in the JAK2, CALR, or MPL genes in >90% of these cases.15 Therefore, mutations in these genes are a major diagnostic criterion for BCR-ABL1-negative MPNs in the 2016 WHO classification of myeloid neoplasms.9 Detecting the presence of these genetic alterations helps in disease diagnosis, which helps prognosis, treatment formulation, and patient management.16 Various molecular techniques like qPCR, allele-specific PCR, fragment length analysis, and capillary sequencing have been used for the detection of mutations in these genes individually or sequentially in a reflex manner. Use of multiple techniques for the molecular diagnosis of MPNs is not only cumbersome and time-consuming, but also limited by the techniques’ inability to detect rare, novel, and low-depth mutations, all of which are generally observed in MPNs. The majority of these assays are targeted to detect canonical mutations only. Moreover, assays like the ARMS PCRs and qPCRs are targeted to detect a specific mutation at a fixed position and may not detect rare or novel mutations within the same region. Therefore, our NGS-based assay was designed to detect both canonical and non-canonical mutations in the reported MPN hotspots—JAK2 exons 12 and 14, CALR exon 9, and MPL exon 10.
In a recent study published by Seon Young Kim et al,17 the authors analyzed the JAK2, CALR, and MPL genes of 199 patients with MPN. In this study, the authors observed that the JAK2 V617F mutation was the most frequently encountered mutation (67.3%), followed by mutations in the CALR (12.6%), and MPL (3.5%) genes. The JAK2 V617F mutation has been reported as the most frequently encountered mutation in patients with MPN in other studies as well.18, 19, 20 Moreover, a study in an Indian cohort of 50 PMF patients showed that 64% of the patients harbored the JAK2 V617F mutation, 26% of the patients harbored mutations in the CALR gene, and 2% in the MPL gene.10 Of the 274 samples tested, we observed the JAK2 V617F exon 14 mutation in only 48 patients (17.5%). One plausible explanation for the low percentage of samples harboring the JAK2 V617F mutation in our cohort could be that for most of the samples referred to our lab, the hematologists performed the JAK2 V617F mutation testing elsewhere; only JAK2 V617F negative samples were referred to our lab to test for other driver mutations by the targeted NGS assay.
We detected a rare pathogenic indel, c.1614_1616delinsATT (H538Q, K539L), in JAK2 exon 12 (Figure 3) in a 28-year-old male patient who presented with portal vein and splenic vein thrombosis with splenomegaly and high Hb levels (17.1 gm/dl). This rare pathogenic indel variant in exon 12 results in an amino acid change from histidine to glutamine at position 538 (H538Q) and from lysine to leucine at position 539 (K539L). It is generally observed in cases of polycythemia vera and primary erythrocytosis with low serum erythropoietin levels. Of the 274 samples tested, 193 patients (70.2%) did not exhibit any driver mutations in the JAK2, CALR, and MPL genes. The predominance of triple-negative MPNs (over JAK2 V617F mutation) in our study could be multifold. It is possible that some of the patients clinically suspected of having an MPN did not have a clonal myeloproliferation and the clinical symptoms such as the increase in Hb, WBC, and platelet counts was owing to secondary or reactive conditions. Secondly, the clinical MPN symptoms could be caused by mutations in genes other than the three targeted in this study. Grinfield et al sequenced the exonic regions of 69 genes implicated in myeloid cancer from more than 2000 patients. Notably, their study showed that genes other than JAK2, MPL, and CALR were implicated in more than 55% of the cases.21 A plethora of somatic mutations in genes implicated in MPNs and other myeloid malignancies have recently come to the fore as a result of extensive NGS-based studies.15 Importantly, these mutations have also been reported in cases of clonal hematopoiesis of indeterminate potential (CHIP), which makes these genes ideal candidates for a targeted NGS assay capable of detecting low-depth mutations. These genes have diverse functions such as DNA methylation (IDH1/2;22) histone modification (ASXL1; 23), mRNA splicing (SRSF2; 24), and cell signaling (NRAS/KRAS; 25).
An NGS assay helps ascertain the depth of a particular mutation, which helps the clinician understand the disease burden in terms of percentage of cells that are affected by the mutation. Our NGS assay can detect mutations at depths as low as 0.5%12 whereas capillary sequencing can detect mutations reliably at depths of only 15-20% and above.11 Higher sequencing depth increases the accuracy of a sequencing assay for detecting low allele burden mutations.26 Almost half the samples harboring mutations in the JAK2 gene exhibited mutation depths greater than 50%; however, more than 90% of the samples harboring mutations in the CALR gene exhibited mutation depths lower than 20%. Moreover, both the samples harboring mutations in the MPL gene exhibited mutation depth lower than 20%. This highlights the diagnostic utility of an NGS-based assay as it can help in the diagnosis of cases where the patients are harboring driver mutations with low mutation depths.
A major limitation of our study is that we do not have the clinical data for all the samples tested by our in-house MPN NGS. Integration of the clinical data with the NGS results would give us a much deeper insight into the prevalence of driver mutations in the JAK2, CALR, and MPL genes in individuals with different MPN subtypes in the Indian population. However, here, we aimed to demonstrate the utility of our NGS assay for the molecular analysis of the three most-commonly implicated genes in Indian patients with suspected MPN. This panel has great utility in a resource-limited setting with financial constraints where patients cannot afford elaborate panels. Additionally, our results indicate the need for a larger panel that takes into account other genes that have been associated with MPNs.
Article DOI
