Characterization of 35 new cases with four different MPLW515 mutations and essential thrombocytosis or primary myelofibrosis

Recently, mutations of MPL, the gene coding for the thrombopoetin receptor, were demonstrated in ~5% of cases of primary myelofibrosis (PMF) and in ~1% of all cases of essential thrombocytosis (ET).1,2 They represent gain-of-function mutations that confer constitutive activation of the JAK-STAT pathway like JAK2V617F.1,2 Two different amino acid exchanges of codon W515 resulting in a tryptophane to leucine (W515L) or lysine (W515K) were described. So far, W515 mutations have been found in ET and PMF, but were never detected in polycythemia vera (PV). Most cases had wild type JAK2V617.1,3 To evaluate the MPLW515 mutations as markers for routine diagnostics of JAK2V617 unmutated myeloproliferative neoplasms (MPN), we performed analyses for MPLW515 mutations in a total of 869 selected MPN patients (399 males; 470 females; 12.2–90.3 years; median 60.5 years) from January 2006 to December 2007. The patients seleted presented ET or suspected ET due to high thrombocyte counts (n=356), or PMF (n=193). In addition, 269 unclassified MPN and 51 PV were analyzed. There was a strong selection towards ET patients with JAK2V617 wild type (324/356) as we were mainly interested in further genetic characterization of this subgroup. Only 32 JAK2V617F mutated ET and 89 JAK2V617F mutated PMF were investigated to look for potential double mutations. Analysis for the MPLW515 mutation status was performed on peripheral blood (519 samples) or bone marrow (350 samples) by a melting curve based LightCycler assay with primers spanning W515 as previously described.4 Cases with altered melting curve patterns were further analyzed by sequencing (Figure 1). Sensitivity of the assay was estimated by a limiting dilution assay (cDNA with homozygous MPLW515K in MPLW515wt cDNA) and was at least 5% (Online Supplementary Figure 1). Analysis for JAK2V617F was performed as previously described.5 Cases were further evaluated by cytomorphology, cyto-chemistry, histopathology, and cytogenetics/FISH. The classification of disorders followed WHO criteria.6

Table 1.Summary of patients characteristics.

In total, 35 MPLW515 mutations were detected in the 869 selected patients. A detailed description of these MPLW515 mutated patients is given in Table 1. In the total cohort with JAK2V617wt ET, any MPLW515 mutation was detected in 19 out of 324 patients (5.9%). In 104 JAK2V617wt PMF, a total of 10 MPLW515 mutations was detected (9.6%). In contrast, in the JAK2V617F mutated cases (32 ET; 89 PMF), no MPLW515 mutation was detected. Sequencing of the 35 MPLW515 mutations revealed four different MPLW515 subtypes: 20 patients had a W515L mutation, 13 a W515K, one case showed a so far not described W515A mutation leading to a tryptophane to alanine exchange, and another case a novel W515R associated to a tryptophane to arginine exchange (Figure 1). Although the functional relevance of these two new mutations still has to be evaluated, it has to be hypothesized that the replacement of a large amino acid by a smaller one probably alters the protein structure in both novel mutation subtypes.

Figure 1.(A) The novel MPLW515A mutation in case 1. This case revealed an approximately 0.2 mutation to wild type (wt) level. On the left the melting curve assay that detects the mutation. On the right further characterization by sequencing compared to a wt allele. (B) The novel MPLW515R mutation in case 28. On the left the melting curve assay that detects the mutation. On the right further characterization by sequencing compared to a wt allele. In addition to the W515R mutation this case has a silent mutation at AA position R515 due to g>a exchange.

Mutation/wild type ratios of greater than 1.0 indicated that at least some cells in the respective patient showed loss of the wildtype allele (LOH). Such high mutation ratios were detected in 13/35 (37%) mutated cases. Eight of these cases with high ratios were at advanced stage with a disease history of 2–9 years. Mutation ratios greater than 1% for the W515 were more frequent in PMF/s-AML following PMF (6/8 cases; 75%) when compared to ET (7/26; 27%) (Online Supplementary Table S1a). The mutation/wild type ratios in the remaining 22 patients were between 0.1 and 0.9% (median: 0.3%) (Online Supplementary Table S1b). Based on the applied method, cells with LOH could not be excluded in these low level cases because unapparent homozygousity based on dilution of homozygous cells with wild type cells may be present. However, the low level cases at least had less cells with LOH and thus the ratios may be important. This allows the hypothesis that higher proportions of W515 mutated alleles in total could indicate progression of disease. High mutation ratios were more frequent in the W515K (9/13; 69%) than in W515L (3/20; 15%) (p=0.034) corresponding to the results of Vannucchi et al.7 and Beer et al.8 Thus, the W515K mutation seems more often associated with loss of the wildtype allele. Karyotypes were available in 12 MPLW515 mutated cases. Eleven had a normal karyotype; one case had a del(5)(q14q34) and a del(13)(q12q22). It is remarkable that at least at the microscopic level, no LOH of chromosome 1p, where MPL is located, was detectable. This issue has to be investigated with more sophisticated techniques like SNP-array analyses.

The frequency of the MPLW515 mutation in our cohort corresponded to previous studies with 5.3% in the JAK2V617wt ET and 9.6% in JAK2V617wt PMF.1,2 In contrast to previous findings, in our small cohort of 121 JAK2V617F mutated patients with ET and PMF there was no case with an MPLW515, whereas others found such a coexistence in up to 22% of MPL mutated MF cases.1,3,9

Finally, the W515 mutations have so far been identified in ET and PMF only.1 Based on the new potential of the MPLW515 mutation in diagnostics, here one case (n. 35, Table 1) which had previously been classified as CMML probably has to be reclassified as ET due to thrombocytosis and the W515L mutation. As mutation analysis for MPLW515 mutations is easy and fast to perform, this case is a good example of how the respective mutation is now of potential help in routine diagnostics to reclassify suspected myeloproliferative diseases and discriminate them from reactive disorders.

Footnotes

• The online version of this article contains a supplementary appendix