Allogeneic stem cell transplantation (aSCT) is a curative treatment approach for patient with myelofibrosis. Monitoring patients after aSCT using molecular disease markers has become a valid tool to guide adoptive immunotherapy with donor lymphocyte infusion in chronic myeloid leukemia and other diseases.1 In particular, myelofibrosis-specific mutations such as JAK2V617F (~50% of patients) and MPLW515L and MPLW515K (~5%) have successfully been used for post-aSCT minimal residual disease (MRD) monitoring.2,3,4,5

The recently discovered somatically acquired calreticulin (CALR) mutations in ~30% of all and 80% of the Janus kinase 2 (JAK2)- and myeloproliferative leukemia protein (MPL)-negative myelofibrosis patients6,7,8 provide new diagnostic markers, which can also be used as indicators of MRD following curative treatment with aSCT.

More than 80% of CALR mutations are of one of two types: type-1 variants result from a 52-bp deletion and produce the protein change p.L367fs*46, and type-2 variants are caused by a 5-bp insertion and produce the protein change p.K385fs*47. Whereas Sanger as well as next-generation sequencing readily allow detection of CALR mutations in newly diagnosed patients, their applicability for MRD detection is limited. Fragment-analysis and high-resolution melting PCR are alternative methods with higher sensitivities, which, however, do not exceed 0.1%.9 Initial studies on monitoring CALR mutations post-aSCT reported rather low sensitivities (1%) insufficient to detect minimal molecular disease.5

Digital PCR (dPCR) is a new PCR method combining the sensitivity of quantitative PCR (qPCR) with excellent accuracy.10,11,12 We here developed a duplex-dPCR assay detecting the CALR type-2 mutation in combination with its wild-type allele (for details, Supplementary Information). To address sensitivity and specificity of the novel dPCR assay, we first tested it on buffy-coat cells from healthy donors and on serial dilutions of a new UKE-113 derived cell line harboring one copy of the CALR type-2 mutation (UKE-1CALRt2). UKE-1CALRt2 was established by transducing UKE-1 cells with a lentiviral LeGO-G/puro vector14 containing the p.K385fs*47 region of mutated CALR together with enhanced green fluorescent protein and puromycin-resistance genes, followed by single-cell sorting. We applied 120 ng EcoRI-restricted genomic DNA of buffy coat and UKE-1CALR2-dilution samples for dPCR.

Out of 106 tested CALR-negative samples (including 48 buffy coats from healthy donors) 94 (89%) were negative in dPCR. The remaining 12 samples showed a mean of 2.2 (range 1.4–4.0) positive events in the 20-μL reaction mix. Based thereon we defined samples with <5 CALRt2 signals per reaction (<0.25 alleles per μL) as negative. For the amount of genomic DNA tested (120 ng) this corresponds to a theoretical detection limit of ~1 in 5000 cells. Well in agreement with this prediction, we readily detected one UKE-1CALRt2 in 5000 buffy-coat cells (0.02%) using our dilution samples (Figure 1a). Importantly, no-template controls included in all PCR runs were always negative.

Figure 1
figure 1

High sensitivity of CALRt2-dPCR based MRD diagnostics. (a) A novel cell line harboring one CALR (K385fs*47) allele was mixed at different ratios with MNC from a healthy donor. As evident, K385fs*-positive cells can be detected at concentrations as low as 0.02%. (b, c) Kinetics of hematopoietic chimerism (HC) and MRD after aSCT for patients 7 (b) and 1 (c). HC was assessed by qPCR, MRD by the new dPCR technique detecting the CALR (K385fs*47) mutation. As illustrated, mutation-specific dPCR results inversely correlate with chimerism in blood and bone marrow. In both patients, CMR was only achieved after therapeutic intervention.

Full size image

Applying our new technique, we next performed MRD analysis in CALRt2+ patients after aSCT and compared results with respective qPCR data. Out of 143 patients with myelofibrosis who underwent allogeneic SCT, 92 were JAK2V617 positive, four MPL positive and 35 CALR positive. Of these 35 patients, 21 harbored the CALR type-1 and eight the CALR type-2 mutation. For seven of those eight patients we were able to apply both qPCR and dPCR for MRD monitoring after SCT (Table 1). Notably, 10 post-SCT peripheral-blood (PB) samples were tested CALRt2-positive by dPCR, whereas only two of them were also found positive by qPCR, albeit at strongly reduced percentages. In one patient (#3), dPCRs were negative at all three time points after transplantation and he has remained in complete molecular remission for 3 years now. In two patients each, dPCR has turned negative at 100 and 180 days post SCT and remained below detection level since; again, all of them have since been in continued complete molecular remission. The two other patients showed different courses. Patient 7 (Figure 1b) was only CALRt2-negative at one time point (at day +100), and showed increasing numbers of CALRt2-positive cells with decreasing chimerism thereafter. In this patient immunosuppression was ceased starting ~5 months post SCT; he has subsequently converted to CALRt2-negativity and full donor chimerism. Patient 1 (Figure 1c) became CALRt2-negative early after SCT. The patient turned CALRt2-positive in the bone marrow (BM) by dPCR at ~18 months and relapsed at 28 months after transplantation. Remarkably, by qPCR the CALRt2 mutation was only detected 1 month before relapse. It is also of note that analysis of bone marrow and/or CD34+ cells allowed more sensitive assessment of molecular relapses. Immunosuppression was tapered and the patient achieved transient improvement in CALRt2 and chimerism levels (months 36–48) followed by another increase in malignant-cell numbers. The patient received a second allograft at 58 months post SCT and has thereafter been in complete molecular remission.

Table 1 HC and CALRt2 positivity measured by dPCR and qPCR in peripheral blood after SCT
Full size table

In conclusion, our data indicate that the new CALR-type-2-mutation-specific dPCR assay combines excellent accuracy with high sensitivity thus allowing the monitoring of deep molecular remission and the early detection of MRD in relapsing patients with myelofibrosis after stem cell transplantation.