Evaluation of the Interest of BRAF –V600E Mutation Detection By Arms-qPCR in Hairy Cell Leukemia Diagnosis and Treatment Monitoring

Background: The discovery of BRAF V600E (Tiacci et al. N.Eng J Med, 2011), has introduced molecular biology in the management of Hairy cell leukemia (HCL). Various techniques for BRAF detection with a specificity of 100% were developed. According to literature, BRAF mutation has been reported in one case of CLL and one B-prolymphocytic leukemia (Langabeer et al. Leukemia research, 2012).

The development of BRAF inhibitors for refractory HCL to purine nucleoside analogues (PNA) renders the detection of BRAF mutation indispensable for the diagnosis and monitoring of the minimal residual disease (MRD).

Objectives: To test a quantitative PCR on HCL patients at diagnosis, during relapse and at MRD monitoring compared to flow cytometry (FCM), then evaluate its usage on peripheral blood (PB) versus bone marrow (BM) sampling.

Methods: We developed a relative quantitative amplification refractory mutation PCR technique (ARMS-qPCR). A retrospective study on a total of 99 samples between 1998 and 2014 was conducted. These samples were previously analyzed by a 4 colors FCM.

38 patients with HCL were tested at diagnosis (21 samples from PB and 17 from BM). 36 patients with other hematologic malignancies were studied, 5 other hairy cell proliferations (3 HCL-Variant (HCL-V) and 2 splenic lymphocyte villous lymphomas SLVL) and 31 non hairy cell proliferations (13 chronic lymphocytic leukemias (CLL), 13 atypical CLL, 3 large B-cell non Hodgkin lymphomas, 2 mantle cell lymphomas, 1 marginal zone lymphoma, 1 acute myeloid leukemia, 1 mastocytosis).

Subsequently, we studied 14 patients in relapse (8 samples from PB, 5 from BM and 1 from a diaphragm nodule). Among these patients, 9 had received PNA, 3 interferon (IFN), 1 PNA + rituximab and 1 patient with unknown treatment.

We monitored 11 patients for MRD (6 samples from PB and 5 from BM), 6 were treated with PNA, 1 with IFN and 1 with PNA + rituximab and 3 patients with unknown treatment.

Results: The sensitivity of our ARMS-qPcr technique attained 0.001%.

At diagnosis, the tumor cells ranged from 0.5% to 91% in PB and 3% to 73% in BM. All patients diagnosed as HCL by FCM were also detected by our PCR technique. The average mutated BRAF allele was 8% (0.02-18%) at diagnosis in the PB and 6% (2.5 -14%) in BM. The mutation was not detected for any of the patients harboring other hairy cell proliferations.

Of the 36 patients with other hematologic malignancies, no signal was detected except for a weak one for 2 CLL patients with atypical morphology and 1 mixed type CLL/prolymphocytic leukemia (mutated allele: 0.03%, 0.21% and 1% respectively). This tempers the total specificity of BRAF detection in HCL described in earlier publications.

At relapse, tumor cells ranged from 0.3% to 81% in PB and 3% to 30% in BM. The average mutated BRAF allele was 5.4% (0.09% -20%) in PB and 3.4% (0.6% -7.5%) in BM.

Concerning evaluation of MRD, tumor cells ranged from 0% to 3% in PB and 0% to 2% in BM. All patients having malignant cells detected by FCM were detected by ARMS-qPCR. 2 patients were undetectable by both methods, 1 was treated by PNA + rituximab and the treatment was unknown for the other. The average mutated BRAF allele was 0.5% (0% -1%) in PB and 0.3% (0% -1%) in BM. Patients treated by PNA or IFN and tested in our study were positive for the MRD.

DNA was obtained simultaneously from PB and BM for 3 patients (2 diagnosis and 1 relapse). The BRAF mutated allele was similar in PB and BM for 2 patients (4% and 11% in PB, 4% and 14% in BM respectively) and the circulating tumor cells were 60% and 61%. As for the 3^rd patient, a lower mutated allele percentage (0.09%) was detected with 1% of tumor cells in blood, where 4% of the mutated allele was found in BM with 25% of tumor cells.

Conclusion: This ARMs-qPcr allowed the HCL diagnosis of 100% of tested patients and the differential diagnosis with other hairy cell proliferations though a weak signal was detected in 3 atypical CLL cases. Using this PCR for relapse detection and MRD monitoring is as performing as the FCM.

PB sampling which is less invasive than BM puncture seems to be adequate for HCL molecular study. BM aspirate may be considered in case of diagnostic difficulties or very few tumor cells.

All patients treated by PNA or IFN and tested in our study were positive for the MRD suggesting that a complete molecular response is hardly achieved with these drugs. This should be confirmed on a larger number of patients as well with the emerging therapies.