Portal de Periódicos da CAPES

Quantification of the Mutant CALR Allelic Burden by Digital PCR

2015; Elsevier BV; Volume: 18; Issue: 1 Linguagem: Inglês

10.1016/j.jmoldx.2015.07.007

1943-7811

Olivier Mansier, Marina Migeon, A. Saint‐Lézer, Chloé James, Emmanuelle Verger, Marie Robin, Gèrard Socié, Audrey Bidet, François‐Xavier Mahon, Bruno Cassinat, Éric Lippert,

Chronic Myeloid Leukemia Treatments

With the recent discovery of CALR mutations, >80% of patients with myeloproliferative neoplasms carry a phenotype-driving mutation. For JAK2 V617F, the most frequent mutation in myeloproliferative neoplasms, accurate determination of mutational loads is of interest at diagnosis, for phenotypic and prognostic purposes, and during follow-up for minimal residual disease assessment. We developed a digital PCR technique that allowed the accurate determination of CALR allelic burdens for the main mutations (types 1 and 2). Compared with the commonly used fluorescent PCR product analysis, digital PCR is more precise, reproducible, and accurate. Furthermore, this method reached a very high sensitivity. We detected at least 0.025% CALR mutants. It can thus be used for patient characterization at diagnosis and for minimal residual disease monitoring. When applied to patients with primary myelofibrosis who underwent hematopoietic stem cell transplant, the digital PCR detected low levels of minimal residual disease. After negativation of the mutational load in all patients, the disease reappeared at a low level in one patient, preceding hematologic relapse. In conclusion, digital PCR adapted to type 1 and 2 CALR mutations is an inexpensive, highly precise, and sensitive technique suitable for evaluation of myeloproliferative neoplasm patients during follow-up. With the recent discovery of CALR mutations, >80% of patients with myeloproliferative neoplasms carry a phenotype-driving mutation. For JAK2 V617F, the most frequent mutation in myeloproliferative neoplasms, accurate determination of mutational loads is of interest at diagnosis, for phenotypic and prognostic purposes, and during follow-up for minimal residual disease assessment. We developed a digital PCR technique that allowed the accurate determination of CALR allelic burdens for the main mutations (types 1 and 2). Compared with the commonly used fluorescent PCR product analysis, digital PCR is more precise, reproducible, and accurate. Furthermore, this method reached a very high sensitivity. We detected at least 0.025% CALR mutants. It can thus be used for patient characterization at diagnosis and for minimal residual disease monitoring. When applied to patients with primary myelofibrosis who underwent hematopoietic stem cell transplant, the digital PCR detected low levels of minimal residual disease. After negativation of the mutational load in all patients, the disease reappeared at a low level in one patient, preceding hematologic relapse. In conclusion, digital PCR adapted to type 1 and 2 CALR mutations is an inexpensive, highly precise, and sensitive technique suitable for evaluation of myeloproliferative neoplasm patients during follow-up. Philadelphia chromosome–negative myeloproliferative neoplasms (MPNs) are characterized by an excessive proliferation of hematopoietic cells consecutive to the clonal acquisition of a phenotype-driving mutation. More than 95% of patients with polycythemia vera (PV), 60% to 70% with essential thrombocytemia (ET), and 50% to 60% with primary myelofibrosis (PMF) harbor a G1849T (V617F) mutation of the JAK2 gene.1James C. Ugo V. Le Couédic J.-P. Staerk J. Delhommeau F. Lacout C. Garçon L. Raslova H. Berger R. Bennaceur-Griscelli A. Villeval J.L. Constantinescu S.N. Casadevall N. Vainchenker W. A unique clonal JAK2 mutation leading to constitutive signalling causes polycythaemia vera.Nature. 2005; 434: 1144-1148Crossref PubMed Scopus (2909) Google Scholar, 2Lippert E. Boissinot M. Kralovics R. Girodon F. Dobo I. Praloran V. Boiret-Dupré N. Skoda R.C. Hermouet S. The JAK2-V617F mutation is frequently present at diagnosis in patients with essential thrombocythemia and polycythemia vera.Blood. 2006; 108: 1865-1867Crossref PubMed Scopus (214) Google Scholar, 3Tefferi A. Lasho T.L. Huang J. Finke C. Mesa R.A. Li C.Y. Wu W. Hanson C.A. Pardanani A. Low JAK2V617F allele burden in primary myelofibrosis, compared to either a higher allele burden or unmutated status, is associated with inferior overall and leukemia-free survival.Leukemia. 2008; 22: 756-761Crossref PubMed Scopus (197) Google Scholar Many JAK2 V617F–negative PVs carry a mutation in the exon 12 of JAK2,4Scott L.M. Tong W. Levine R.L. Scott M.A. Beer P.A. Stratton M.R. Futreal P.A. Erber W.N. McMullin M.F. Harrison C.N. Warren A.J. Gilliland D.G. Lodish H.F. Green A.R. JAK2 exon 12 mutations in polycythemia vera and idiopathic erythrocytosis.N Engl J Med. 2007; 356: 459-468Crossref PubMed Scopus (1006) Google Scholar whereas the gene encoding for the JAK2-coupled thrombopoietin receptor MPL is found mutated in 5% to 10% of ET or PMF patients.5Pikman Y. Lee B.H. Mercher T. McDowell E. Ebert B.L. Gozo M. Cuker A. Wernig G. Moore S. Galinsky I. DeAngelo D.J. Clark J.J. Lee S.J. Golub T.R. Wadleigh M. Gilliland D.G. Levine R.L. MPLW515L is a novel somatic activating mutation in myelofibrosis with myeloid metaplasia.PLoS Med. 2006; 3: e270Crossref PubMed Scopus (1105) Google Scholar Recently, recurrent mutations in the CALR gene were discovered in most JAK2 and MPL nonmutated ET and PMF patients.6Klampfl T. Gisslinger H. Harutyunyan A.S. Nivarthi H. Rumi E. Milosevic J.D. Them N.C.C. Berg T. Gisslinger B. Pietra D. Chen D. Vladimer G.I. Bagienski K. Milanesi C. Casetti I.C. Sant'Antonio E. Ferretti V. Elena C. Schischlik F. Cleary C. Six M. Schalling M. Schönegger A. Bock C. Malcovati L. Pascutto C. Superti-Furga G. Cazzola M. Kralovics R. Somatic mutations of calreticulin in myeloproliferative neoplasms.N Engl J Med. 2013; 369: 2379-2390Crossref PubMed Scopus (1460) Google Scholar, 7Nangalia J. Massie C.E. Baxter E.J. Nice F.L. Gundem G. Wedge D.C. et al.Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2.N Engl J Med. 2013; 369: 2391-2405Crossref PubMed Scopus (1360) Google Scholar In addition to better understanding the pathophysiology of MPN, the discovery of these mutations has provided very useful clonal markers for the diagnosis of MPN. Further characterization of the diseases has been permitted by the accurate assessment of the mutational burden, especially for the JAK2 V617F mutation. For instance, it has been rapidly recognized that the JAK2 V617F allelic burden is much higher in PV than in ET patients,2Lippert E. Boissinot M. Kralovics R. Girodon F. Dobo I. Praloran V. Boiret-Dupré N. Skoda R.C. Hermouet S. The JAK2-V617F mutation is frequently present at diagnosis in patients with essential thrombocythemia and polycythemia vera.Blood. 2006; 108: 1865-1867Crossref PubMed Scopus (214) Google Scholar most of the latter having mainly heterozygously mutated cells.8Scott L.M. Scott M.A. Campbell P.J. Green A.R. Progenitors homozygous for the V617F mutation occur in most patients with polycythemia vera, but not essential thrombocythemia.Blood. 2006; 108: 2435-2437Crossref PubMed Scopus (228) Google Scholar In addition, different levels of mutant allele burden have been associated with various outcomes,3Tefferi A. Lasho T.L. Huang J. Finke C. Mesa R.A. Li C.Y. Wu W. Hanson C.A. Pardanani A. Low JAK2V617F allele burden in primary myelofibrosis, compared to either a higher allele burden or unmutated status, is associated with inferior overall and leukemia-free survival.Leukemia. 2008; 22: 756-761Crossref PubMed Scopus (197) Google Scholar, 9Vannucchi A.M. Antonioli E. Guglielmelli P. Rambaldi A. Barosi G. Marchioli R. Marfisi R.M. Finazzi G. Guerini V. Fabris F. Randi M.L. De Stefano V. Caberlon S. Tafuri A. Ruggeri M. Specchia G. Liso V. Rossi E. Pogliani E. Gugliotta L. Bosi A. Barbui T. Clinical profile of homozygous JAK2 617V>F mutation in patients with polycythemia vera or essential thrombocythemia.Blood. 2007; 110: 840-846Crossref PubMed Scopus (372) Google Scholar, 10Guglielmelli P. Barosi G. Specchia G. Rambaldi A. Lo Coco F. Antonioli E. Pieri L. Pancrazzi A. Ponziani V. Delaini F. Longo G. Ammatuna E. Liso V. Bosi A. Barbui T. Vannucchi A.M. Identification of patients with poorer survival in primary myelofibrosis based on the burden of JAK2V617F mutated allele.Blood. 2009; 114: 1477-1483Crossref PubMed Scopus (162) Google Scholar emphasizing the need for accurate and reproducible measurements of mutant allelic burdens. Last, these clonal markers can be used for disease burden assessments and monitoring the effect of some treatments. Interferon alpha has thus been found to efficiently decrease JAK2 V617F allelic burden in PV and ET patients, some of whom eventually reach a level undetectable by current techniques.11Kiladjian J.-J. Cassinat B. Chevret S. Turlure P. Cambier N. Roussel M. Bellucci S. Grandchamp B. Chomienne C. Fenaux P. Pegylated interferon-alfa-2a induces complete hematologic and molecular responses with low toxicity in polycythemia vera.Blood. 2008; 112: 3065-3072Crossref PubMed Scopus (457) Google Scholar Hematopoietic stem cell transplant (HSCT) is equally able to strongly decrease JAK2 V617F levels, which can be used as a minimal residual disease (MRD) marker, the persistence of which predicts a higher risk of relapse.12Kröger N. Badbaran A. Holler E. Hahn J. Kobbe G. Bornhäuser M. Reiter A. Zabelina T. Zander A.R. Fehse B. Monitoring of the JAK2-V617F mutation by highly sensitive quantitative real-time PCR after allogeneic stem cell transplantation in patients with myelofibrosis.Blood. 2007; 109: 1316-1321Crossref PubMed Scopus (133) Google Scholar In addition, the persistence or reappearance of the mutation can guide the modulation of the immunotherapy after HSCT to complete the molecular response.13Kröger N. Alchalby H. Klyuchnikov E. Badbaran A. Hildebrandt Y. Ayuk F. Bacher U. Bock O. Kvasnicka M. Fehse B. Zander A. JAK2-V617F-triggered preemptive and salvage adoptive immunotherapy with donor-lymphocyte infusion in patients with myelofibrosis after allogeneic stem cell transplantation.Blood. 2009; 113: 1866-1868Crossref PubMed Scopus (66) Google Scholar These elements make an accurate sensitive and reproducible technique for evaluation of mutant allelic burden extremely useful for characterization and monitoring of MPN patients. Whereas the JAK2 V617F mutation affects a single nucleotide, mutations in the exon 9 of the CALR gene are more diverse. Most mutations are type 1 (45% to 53%), corresponding to a 52-bp deletion, or type 2 (32% to 41%), consisting of a 5-bp insertion.6Klampfl T. Gisslinger H. Harutyunyan A.S. Nivarthi H. Rumi E. Milosevic J.D. Them N.C.C. Berg T. Gisslinger B. Pietra D. Chen D. Vladimer G.I. Bagienski K. Milanesi C. Casetti I.C. Sant'Antonio E. Ferretti V. Elena C. Schischlik F. Cleary C. Six M. Schalling M. Schönegger A. Bock C. Malcovati L. Pascutto C. Superti-Furga G. Cazzola M. Kralovics R. Somatic mutations of calreticulin in myeloproliferative neoplasms.N Engl J Med. 2013; 369: 2379-2390Crossref PubMed Scopus (1460) Google Scholar, 7Nangalia J. Massie C.E. Baxter E.J. Nice F.L. Gundem G. Wedge D.C. et al.Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2.N Engl J Med. 2013; 369: 2391-2405Crossref PubMed Scopus (1360) Google Scholar The primer-based allele-specific PCRs that were the most reliable for JAK2 V617F mutations14Lippert E. Girodon F. Hammond E. Jelinek J. Reading N.S. Fehse B. Hanlon K. Hermans M. Richard C. Swierczek S. Ugo V. Carillo S. Harrivel V. Marzac C. Pietra D. Sobas M. Mounier M. Migeon M. Ellard S. Kröger N. Herrmann R. Prchal J.T. Skoda R.C. Hermouet S. Concordance of assays designed for the quantification of JAK2V617F: a multicenter study.Haematologica. 2009; 94: 38-45Crossref PubMed Scopus (66) Google Scholar, 15Jovanovic J.V. Ivey A. Vannucchi A.M. Lippert E. Oppliger Leibundgut E. Cassinat B. et al.Establishing optimal quantitative-polymerase chain reaction assays for routine diagnosis and tracking of minimal residual disease in JAK2-V617F-associated myeloproliferative neoplasms: a joint European LeukemiaNet/MPN&MPNr-EuroNet (COST action BM0902) study.Leukemia. 2013; 27: 2032-2039Crossref PubMed Scopus (84) Google Scholar cannot be easily adapted for the quantification of the larger CALR mutations. This approach has been successfully used to quantify type 1 and type 2 mutations with a limit of detection of approximately 1%, a level that is suboptimal for MRD monitoring.16Chi J. Manoloukos M. Pierides C. Nicolaidou V. Nicolaou K. Kleopa M. Vassiliou G. Costeas P. Calreticulin mutations in myeloproliferative neoplasms and new methodology for their detection and monitoring.Ann Hematol. 2015; 94: 399-408Crossref PubMed Scopus (36) Google Scholar However, this approach of allelic discrimination can reach a higher sensitivity in the context of digital PCR (dPCR). This technique relies on the fragmentation of the PCR reactions into very small droplets that are clonally analyzed.17Pinheiro L.B. Coleman V.A. Hindson C.M. Herrmann J. Hindson B.J. Bhat S. Emslie K.R. Evaluation of a droplet digital polymerase chain reaction format for DNA copy number quantification.Anal Chem. 2012; 84: 1003-1011Crossref PubMed Scopus (714) Google Scholar Thus, the quantification of the PCR results is based on the number of droplets positive for amplification of the wild-type or the mutant studied allele. This results in an absolute reliable and sensitive assessment of the mutational load. Because all CALR mutations imply size variations (deletions and/or insertions), their detection and relative quantification are frequently performed using PCR fragment length analysis. These are typically achieved by capillary electrophoresis of end point PCR products and integration of the area under the peaks of the electropherograms. We compared the performances of fragment analysis to those of allelic discrimination using dPCR. We found that, even though both techniques are appropriate for routine analyses, the latter has a higher accuracy and lower detection threshold, making it a better tool for both characterization of patients and MRD monitoring. DNA was obtained from granulocytes of patients with myeloproliferative disorders explored for the CALR mutation at the hematology laboratory of the CHU de Bordeaux or patients included in the JAKALLO clinical trial (http://www.clinicaltrials.gov; accession number NCT01795677). All patients had provided informed written consent in accordance with the Declaration of Helsinki for the use of the remaining DNA for investigational purposes. DNA was extracted using the QIAamp Blood DNA minikit according to the manufacturer's instructions (Qiagen, Hilden, Germany). As a means of developing standards for assessing the precision of the method, a sequence corresponding to exon 9 of CALR was amplified from patients with wild-type, type 1, or type 2 CALR mutations with primers CCF and CCR (Table 1) at 500 nmol/L and a mix that contained 1.5 mmol/L MgCl2, 0.2 mmol/L dNTP, 1.25U GoTaq, and the provider's specific buffer (Promega France, Charbonnieres, France), then cloned into the TopoTA vector (Life Technologies, Carlsbad, CA). The plasmid was quantified by an Evagreen-based ddPCR using primers specific for the plasmid backbone (TOPOR and TOPOF) (Table 1) at 130 nmol/L and the ddPCR QX200 Evagreen supermix. The PCR program was as follows: denaturation for 5 minutes at 95°C, followed by 40 cycles of 30-second denaturation (95°C) and 1-minute annealing/extension (60°C), then 5 minutes at 4°C and 5 minutes at 90°C. Total CALR gene copy numbers were also measured in control normal DNA with ddPCR using the CFF and CCR primers following the same protocol.Table 1Primers and ProbesApplicationNameSequenceCALR cloning, sequencingCCF5′-ATGCTAGCCTGCCGGAGGGTCGTTTTA-3′CCR5′-ATACGCGTAGAGACATTATTTGGCGCGG-3′Evagreen plasmid quantificationTopoF5′-ACACGTAGAAAGCCAGTCCG-3′TopoR5′-TAAGCCCACTGCAAGCTACC-3′Fragment analysisCFF5′-GAGGTGTGTGCTCTGCCT-3′CFR5′-AGAGACATTATTTGGCGCGG-3′dPCR type 1CD1F5′-CAGAGAAACAAATGAAGGACAAAC-3′CD1R5′-GGGACATCTTCCTCCTCATC-3′CD1PW∗Probes.5′-GGACGAGGAGCAGAGGCTTAAGGAGGA-3′CD1PM∗Probes.5′-GGACGAGGAGCAGAGGACAAGGAGGA-3′dPCR type 2CD2F5′-TGAAGGACAAACAGGACGAG-3′CD2R5′-CCGGGGACATCTTCCTCC-3′CD2PW∗Probes.5′-AGGCAGAGGACAATTGTCGGAGGATGATGA-3′CD2PM∗Probes.5′-GAGGAGGCAGAGGACAAGGAGGATGATGA-3′F, forward; M, mutant; dPCR, digital PCR; R, reverse; W, wild type.∗ Probes. Open table in a new tab F, forward; M, mutant; dPCR, digital PCR; R, reverse; W, wild type. For fragment analyses, a 298-bp PCR product was obtained for wild-type CALR after 30 to 35 cycles of amplification using 25 to 75 ng of genomic DNA and the primers CFF and CFR (Table 1), the latter 5′-labeled with fluorescein. The PCR mix using GoTaq was as described above. The PCR program consisted of a 2-minute denaturation at 95°C, 30 to 35 cycles of 45 seconds at 95°C, 1 minute at 60°C, and 2 minutes at 72°C followed by a final extension of 5 minutes at 72°C. After a 1:50 dilution of the PCR products into water, the peaks of wild-type and mutant CALR were resolved on a 3130 sequencer (ABI; Thermo Fisher Scientific, Waltham, MA), and the ratios of mutant to total CALR were determined by measuring the areas under the peaks. For the ddPCR, 25 to 250 ng of genomic DNA were mixed with primers and probes specific for either type 1 or type 2 mutation. In each case, the wild-type probes (CD1PW and CD2PW) (Table 1) were 5′-labeled with HEX (6-carboxy-2,4,4,5,7,7-hexachlorofluorescein succinimidyl ester) and the mutant probes (CD1PM and CD2PM) with FAM (6-carboxyfluorescein); all had a 3′-BHQ quencher. The mix contained primers at 900 nmol/L each: wild-type and mutant probes at 250 nmol/L and the droplet ddPCR supermix for Probes (Biorad, Pleasanton, CA). The emulsion was prepared with the QX-100 (Biorad). The amplification program was as follows: 10-minute denaturation at 95°C, followed by 40 cycles of 30 seconds at 94°C, 1 minute at 63°C (type 1) or 64°C (type 2), and inactivation of 10 minutes at 98°C. The number of droplets positive for wild-type or mutant CALR was determined on the QX-200 droplet reader (Biorad) using the Quantasoft software version 1.5 (Biorad). A total of 10,000 to 18,000 droplets of 0.85 nL were analyzed in each well. Limit of quantification assessed using the blank samples was determined as three times the SD of CALR allelic burden assessed on 15 healthy control samples. Mean bias was determined as the mean difference between allelic burdens assessed by both evaluated techniques, and statistical significance was determined by Deming Regression using GraphPad Prism software version 5.0c (GraphPad Software, San Diego, CA). Differences between techniques on patients' DNA were assessed by the Wilcoxon matched-pairs signed rank test. To assess the accuracy and detection limit of fragment analysis, genomic DNAs from patients with 50% CALR mutant allelic burden (assessed by fragment length analysis and ddPCR) were diluted into DNA from healthy controls. With the use of 75 ng of matrix DNA, dilutions up to 1:16 for type 1 mutations and 1:32 for type 2 still produced a quantifiable peak, indicating a sensitivity of the technique of approximately 3% and 1.5%, respectively (Figure 1). At lower values, the mutant peaks were inconsistently detected. Interrun coefficients of variation varied from 1.7% at 50% allelic burden to 27.5% for lower values with type 1 mutants and from 6% at 50% allelic burden to 26.6% for lower values of type 2 mutants. However, with the use of these conditions optimized for the detection of low quantities of mutant DNA (75 ng of DNA, 35 PCR cycles), the measured allelic burdens were systematically overestimated by a factor as high as 145% for low allelic burdens when compared with theoretical values expected from the dilution factor, suggesting that the technique is suboptimal for accurate allelic burden estimation. To avail ourselves of a method able to both reliably quantify CALR mutant allelic burden and reach a lower limit of detection, we developed a ddPCR technique. The analysis of the serial DNA dilutions described above indicated that this technique is more sensitive than the fragment analysis. With the use of 75 ng of DNA, samples with 0.025% mutant CALR were systematically detected with a limit of quantification of 0.014% for type 1. The detection limit was equally low (0.025%) for type 2, with a limit of detection of 0% (no unspecific positivity) (Figure 2). In these conditions, the number of target copies per partition ranged from 1.34 to 2.5. To improve the sensitivity of the technique, up to 250 ng of genomic DNA per well was analyzed, but this resulted in an increased background with up to 1% false-positive droplets. Interrun coefficients of variation varied from 0.06% to 9.6% for type 1 and from 0.07% to 18.9% for type 2 mutants (depending on the level of allelic burden considered). The observed results were perfectly superimposable with theoretical values expected from the dilution process, suggesting that ddPCR quantification is more accurate than that obtained with fragment analysis. Bland-Altman representations comparing ddPCR and fragment analysis with either dilutions of mutant into normal DNA or primary patients' samples with various levels of CALR mutants confirmed that fragment analysis systematically overestimated values between 1.5% and 35% (Figure 3) (P < 0.0001 for the difference between the two techniques).Figure 3Comparison of allelic burdens determined by fragment analysis and droplet digital PCR (ddPCR) on patients' samples. Allelic burdens of CALR mutations were assessed by fragment analysis and ddPCR. A: The results obtained by the two techniques using primary patients' samples (P < 0.0001, Wilcoxon matched-pairs signed rank test). Bland and Altman representation reveals the bias linked to fragment analysis compared to ddPCR with primary patients' samples (B) or DNA dilutions for type 1 (C) and type 2 (D) mutants.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Depending on the PCR process, results obtained by fragment analysis can be different from those observed with ddPCR. However, exact evaluation of the allelic burden is critical for comparison of results generated by different laboratories. To ensure that the ddPCR values were the most reliable, defined numbers of copies of mutant plasmid DNA were diluted into equally known numbers of wild-type CALR genomic DNA copies, both determined by non–allele-specific ddPCR using Evagreen. Theoretical values were strictly superimposable with observed values, confirming the accuracy of the digital allelic discrimination (mean bias, −0.44% and −0.50% for type 1 and type 2 mutations, respectively; nonsignificant difference by Deming regression) (Figure 4). In contrast, we observed a systematic overestimation of values 50% for both mutant types when using fragment analysis (mean bias, 0.55% and 2.2% for type 1 and type 2 mutations, respectively, with a significant difference determined by Deming regression). To explain this, we reasoned that a saturation of the PCR and/or the peak signals on electrophoresis may explain this discrepancy. To limit these effects, the numbers of PCR cycles was decreased to 25 or 30, and the dilutions of the PCR products were adjusted to obtain unsaturated peaks. These conditions did not improve significantly the accuracy of the quantification (Figure 5) but altered the limit of detection with dilutions at 3% rarely detectable, especially after 25 cycles of PCR.Figure 5Fragment analysis accuracy in nonsaturating conditions. Samples that contained known levels of mutant CALR were prepared by mixing plasmid DNAs with known copy numbers of mutant CALR and normal genomic DNA with known amounts of wild-type CALR, both independently determined by absolute quantification using a non–allele-specific Evagreen-based ddPCR. Allelic burden was assessed for each dilution with fragment analysis after different PCR cycle numbers and different dilutions of the PCR product.View Large Image Figure ViewerDownload Hi-res image Download (PPT) DNA obtained from granulocytes of patients before and after having received an allogeneic HSCT for primary myelofibrosis were evaluated with the allele-specific ddPCR technique (Figure 6). The CALR allelic burden markedly decreased in all four tested patients. In all patients for whom a sample was harvested in the first month after the transplant (unique patient numbers 2, 3, and 4), a minimal signal was detected between 0.03% and 1.1%. Two of them (unique patient numbers 3 and 4) achieved a complete molecular response later. For UPN 1, first tested 4 months after the transplant, the first control had no residual disease. However, 2 months later, the disease reappeared (0.03%), which was indicative of a relapse that was confirmed on the following assessment. Beside the progress in knowledge of MPN pathophysiology, the discovery of CALR mutations has provided a very useful diagnostic marker for patients lacking JAK2 or MPL mutations. Among the different techniques used for their detection, some do not assess the allelic burden (high-resolution melting curve analyses, Sanger sequencing), yet accurate allelic burden assessment in MPN patients can be of interest. This has been clearly shown for the JAK2 V617F mutation, the burden of which has been associated with specific biological features and evolutive risks. Today, only a few studies have evaluated the impact of CALR mutation burden on the disease phenotype.18Cabagnols X. Defour J.P. Ugo V. Ianotto J.C. Mossuz P. Mondet J. Girodon F. Alexandre J.H. Mansier O. Viallard J.F. Lippert E. Murati A. Mozziconacci M.J. Saussoy P. Vekemans M.C. Knoops L. Pasquier F. Ribrag V. Solary E. Plo I. Constantinescu S.N. Casadevall N. Vainchenker W. Marzac C. Bluteau O. Differential association of calreticulin type 1 and type 2 mutations with myelofibrosis and essential thrombocytemia: relevance for disease evolution.Leukemia. 2015; 29: 249-252Crossref PubMed Scopus (78) Google Scholar, 19Andrikovics H. Krahling T. Balassa K. Halm G. Bors A. Koszarska M. Batai A. Dolgos J. Csomor J. Egyed M. Sipos A. Remenyi P. Tordai A. Masszi T. Distinct clinical characteristics of myeloproliferative neoplasms with calreticulin mutations.Haematologica. 2014; 99: 1184-1190Crossref PubMed Scopus (82) Google Scholar These studies indicate that higher burdens correlate with anomalies of the blood cell counts (higher platelet and leukocyte counts, lower hemoglobin level) and with more advanced stages of their disease. Determining allelic burdens can also be useful when monitoring patients under treatment. JAK2 V617F can decrease and even become undetectable after treatment with interferon alpha or allogeneic stem cell transplant. This seems to also be the case for CALR mutations.20Cassinat B. Verger E. Kiladjian J.-J. Interferon alfa therapy in CALR-mutated essential thrombocythemia.N Engl J Med. 2014; 371: 188-189Crossref PubMed Scopus (58) Google Scholar, 21Haslam K. Langabeer S.E. Molloy K. McMullin M.F. Conneally E. Assessment of CALR mutations in myelofibrosis patients, post-allogeneic stem cell transplantation.Br J Haematol. 2014; 166: 800-802Crossref PubMed Scopus (25) Google Scholar In most studies of CALR mutations, the allelic burden was monitored by fragment length analysis. This technique has frequently been used for the purpose of quantifying relative amounts of size mutants compared with their wild-type allele. Indeed, it is simple, inexpensive, and relatively easy to develop. However, in our hands, care must be taken not to saturate fluorescence signals. Even in these conditions, the quantification accuracy remains suboptimal; moreover, the limit of detection of the technique is rather high, approximately 3%. This may not be an issue for routine screening because most patients with CALR mutations have ratios of 40% to 50% and rarely 1% negatively affects evolution.26Lange T. Edelmann A. Siebolts U. Krahl R. Nehring C. Jäkel N. Cross M. Maier J. Niederwieser D. Wickenhauser C. JAK2 p.V617F allele burden in myeloproliferative neoplasms one month after allogeneic stem cell transplantation significantly predicts outcome and risk of relapse.Haematologica. 2013; 98: 722-728Crossref PubMed Scopus (52) Google Scholar To determine whether identical results could be observed for CALR mutations, we assessed CALR allelic burdens in patients before and after allogeneic SCT for myelofibrosis. As was observed with JAK2 V617F, we noticed a marked decrease followed by a negativation of CALR mutational load in most patients.21Haslam K. Langabeer S.E. Molloy K. McMullin M.F. Conneally E. Assessment of CALR mutations in myelofibrosis patients, post-allogeneic stem cell transplantation.Br J Haematol. 2014; 166: 800-802Crossref PubMed Scopus (25) Google Scholar In one of them, the reappearance of detectable levels of CALR mutation has been detected preceding a relapse. This molecular relapse, which was at low level (0.03%), would have been missed by any other currently available technique. Because this study was retrospective, no therapeutic intervention has been undertaken, but in a prospective setting, one may think of modulating the immune response either by alleviating the immunosuppression or even by injecting donor lymphocytes, as has proven successful in MPN carrying JAK2 V617F mutations.13Kröger N. Alchalby H. Klyuchnikov E. Badbaran A. Hildebrandt Y. Ayuk F. Bacher U. Bock O. Kvasnicka M. Fehse B. Zander A. JAK2-V617F-triggered preemptive and salvage adoptive immunotherapy with donor-lymphocyte infusion in patients with myelofibrosis after allogeneic stem cell transplantation.Blood. 2009; 113: 1866-1868Crossref PubMed Scopus (66) Google Scholar In conclusion, for routine diagnostic purposes, alternative techniques, such as fragment length analysis or even Sanger sequencing, are better suited than ddPCR because they can detect any kind of mutations and low allele burdens are rare at diagnosis. However, ddPCR constitutes, for a relatively modest cost, a sensitive and accurate technique useful for both more accurately characterizing CALR-mutated MPN patients at diagnosis and better monitoring CALR allelic burden in treated patients, with the effect of possibly fine-tuning the therapeutic interventions. We thank Cancéropôle Grand-Sud Ouest and La Ligue Contre le Cancer, Aquitaine-Charentes, for funding this research, the association Parentraides for helping with laboratory equipment, and CRB-Cancer CHU de Bordeaux for providing patient samples.