Molecular detection of BRAF-V600E is superior to flow cytometry for disease evaluation in hairy cell leukaemia

Hairy cell leukaemia (HCL) is a low grade lymphoproliferative B-cell disorder. Diagnosis relies on morphological and immunophenotypic criteria that include the following: CD20, CD22, CD11c, CD25, CD103, tartrate-resistant acid phosphatase and Annexin A1 1. However, diagnostic difficulties can often arise because of a low number of circulating leukemic cells and a ‘dry tap’ on marrow aspiration. The differential diagnosis includes myelofibrosis, splenic marginal zone lymphoma and hairy cell variant (HCLv) that is CD25 negative. Making the correct diagnosis is essential as treatment with nucleoside analogues, such as cladribine and pentostatin, or with interferon or rituximab is usually highly effective in HCL 2.

By using whole exome sequencing, Tiacci et al. in 2011 identified a V600E mutation in exon 15 of the BRAF gene in HCL 3. Remarkably, this mutation was found in all 48 HCL patients tested by Sanger sequencing but was absent in all 195 patients with a variety of other peripheral B-cell lymphomas. This finding was independently verified by high-resolution melting analysis, where the mutation was present in all 48 HCL patients and none of 114 non-HCL patients 4. The BRAF protein encoded on chromosome 7q34 functions as a serine/threonine kinase. The V600E mutation, which is also found in melanoma and other malignancies, causes constitutive activation of the MAP/ERK pathway leading to stimulation of cell growth and survival 5. Targeted inhibition of BRAF by using the inhibitor PLX-4720 (vemurafenib) has been shown to yield haematological remission in a patient with BRAF-V600E mutation-positive refractory HCL 6.

On the basis of these findings, it is likely that BRAF-V600E detection will be incorporated into diagnostic testing for patients suspected of having HCL in the future and indeed may become an obligatory requirement to establish a full diagnosis of HCL. Several techniques have been developed to identify this mutation and form a genetic-based diagnostic test. Techniques developed include allele-specific polymerase chain reaction (PCR) 7, mRNA-based reverse transcription allele-specific PCR 8, melting curve analysis 4 and pyrosequencing 9. However, to our knowledge, none of these tests are currently commercially available. We aimed to validate a commercially available allele-specific quantitative PCR (qPCR) test in a cohort of previously treated HCL patients and to compare the sensitivity and specificity against flow cytometry, considered to be the gold standard method of minimal residual disease (MRD) monitoring in HCL. Our goal was to establish a proof of principle for MRD monitoring by using peripheral blood sampling from patients previously treated for HCL and currently in clinical remission that is based on qPCR detection of BRAF-V600E as a measure of depth of remission. The project was conducted under the Brighton Blood Disorder Research Project and received ethical approval from the East Brighton Local Research Ethics Committee (REC 09/H1107/1).

We investigated a total of 21 samples obtained from 16 patients with previously treated HCL, 4 healthy controls and 1 patient with newly diagnosed HCLv. All diagnoses had been confirmed by flow cytometry of bone marrow samples and exhibited the typical immunophenotype for HCL. All participants gave written informed consent in accordance with the Declaration of Helsinki and their clinical characteristics and previous treatments are summarized in Table 1. Time since initial diagnosis ranged from 6 months to 25 years with a range of time from previous treatment of 6 months to 17 years. All patients had either normal or stable blood counts and were considered to be in clinical remission at the time of entry into the study with no morphologically recognizable hairy cells in the peripheral blood and no clinically detectable splenomegaly on examination. Four patients had notable cytopenias, one because of recent diagnosis and treatment with Cladribine (patient 1) and three others with long-standing stable cytopenias (patients 6, 12 and 13) following previous chemotherapy treatment who were otherwise considered to be in remission from HCL.

Multi-parameter flow cytometry was performed on peripheral blood by using a four colour BD FACS Canto II. Classification of HCL positive cells was based on monotypic surface immunoglobulin bright co-expression of CD20, CD22, CD11c and dim CD103 expression. A minimum of 10 000 events were recorded, and the number of positive events was calculated as a percentage of the mononuclear cell population. For mutation detection, DNA was extracted from peripheral blood by using Qiagen DNeasy kit following the manufacturer's instructions. Samples were blinded and analyzed in triplicate by real time qPCR by using an allele specific amplification system (QUASA System, Primer Design Ltd, Southampton, UK). This method uses a range of novel design approaches to achieve highly specific and sensitive single nucleotide polymorphism quantification without the need for running multiple standard curves. Briefly, the percentage level of the BRAF-V600E mutation was calculated using mutation-specific primer amplified DNA relative to wildtype-specific primer amplified DNA. The combined results of the flow cytometry and qPCR testing are contained in Table 1 with full blood counts at the time of analysis shown in Table 2.

By using flow cytometry, we found that 5 out of 16 HCL samples tested positive for MRD detection defined as ≥0.01% events meeting the co-expression criteria (range 0.01–0.1%). In comparison, by using allele-specific qPCR, we found that 10 out of 16 HCL samples tested positive (range <0.05–0.22%). This included all five samples that were positive by flow cytometry with the remaining six samples testing negative by both techniques. As expected, all four control samples from healthy individuals tested negative by using both methods. Interestingly, the single patient with newly diagnosed HCLv tested negative for BRAF-V600E by qPCR suggesting that this mutation could potentially be used to distinguish HCL from HCLv. The sensitivity of the qPCR test was significantly higher than for flow cytometry, 62.5% vs. 31%. The negative predictive value of the qPCR test was also significantly higher than for flow cytometry, 40% vs. 26.6%. Both methods had 100% specificity and positive predictive value as there were no false positives amongst the normal controls tested.

Of note, one patient (patient 1) was tested at 3 and 6 weeks post initial cladribine treatment. At 3 weeks the mutation was 0.09% by qPCR and 0.02% by flow cytometry but at 6 weeks the mutation was undetectable by both methods. This highlights the value of MRD monitoring by qPCR to assess response to treatment. Interestingly, the patient with the highest mutational percentage by qPCR, and weakly positive by flow cytometry (patient 6), developed pancytopenia because of disease relapse 3 months post-MRD testing. Another patient who tested positive by qPCR but negative by flow cytometry (patient 4) similarly developed neutropenia and thrombocytopenia because of disease relapse 12 months post-MRD testing. Both cases were confirmed by bone marrow biopsy showing infiltration of 25% and 18% HCL cells respectively highlighting the value of BRAF-V600E monitoring for MRD surveillance particularly in patients with new-onset cytopenias. Of interest, patient no. 6 had previously shown clinical refractoriness to cladribine and pentostatin and only partial response with rituximab, whereas patient no. 4 had received interferon and pentostatin. Both individuals would be potential candidates for treatment with the selective BRAF inhibitor, vemurafenib, that has shown impressive responses in the setting of relapsed/refractory HCL 10.

In summary, we conclude that testing for the BRAF-V600E mutation should improve diagnostic accuracy in patients suspected of having a diagnosis of HCL. Moreover, our study suggests that qPCR mutation detection and quantification by using peripheral blood monitoring is a potentially useful method for evaluating disease during post-treatment follow-up. Although this is not currently routine clinical practice for HCL, there is considerable precedent for doing so from other haematological malignancies, notably BCR-ABL in Philadelphia positive chronic myeloid leukaemia and PML-RARα in acute promyelocytic leukaemia. We therefore propose that molecular testing for BRAF-V600E by qPCR by using peripheral blood extracted DNA is both practical and informative in the diagnosis of patients suspected of having HCL and in their long-term follow-up allowing assessment of treatment response and for early detection of disease relapse. Further studies will be required to establish the clinical utility of such an approach.

Conflict of interest

T. Rider, R. Ansell, S. Bastow, R. R. Ghurye, H. Stewart and T. Chevassut declare no competing interests. R. Gover and R. Powell (Primer Design Ltd, UK) declare a commercial interest in the QUASA System as owners of the intellectual property on which the method is based (patent application number 1302112.6).

Author contributions

T. Rider collected samples, performed flow cytometry, extracted DNA and drafted the manuscript. R. Ansell and S. Bastow trained T. Rider in flow cytometry techniques, supervised the flow cytometry and aided data interpretation. R. R. Ghurye assisted with manuscript preparation. H. Stewart trained T. Rider in DNA extraction and contributed to study design, data interpretation and manuscript drafting. T. Chevassut oversaw the design, data collection, data interpretation and manuscript preparation. R. Powell developed the QUASA method for BRAF mutation detection and R. Gover performed the qPCR analysis.