Fusion-harboring mast cells can explain molecular positivity in flow cytometric MRD-negative core-binding factor AML Free

TO THE EDITOR:

Core-binding factor-acute myeloid leukemia (CBF-AML) accounts for approximately 30% of pediatric and 15% of adult AMLs and is characterized by fusions involving RUNX1 or CBFB.^1-4 During normal hematopoiesis, these transcription factors heterodimerize to regulate differentiation.^4,5 In CBF-AML, the fusions RUNX1::RUNX1T1/t(8;21)(q22;q22) (formerly AML1::ETO) or CBFB::MYH11/inv(16), or t(16;16), inhibit normal RUNX1, preventing normal myeloid differentiation. However, these fusions are not sufficient to induce leukemia.^6-11 Reverse transcription polymerase chain reaction (RT-PCR) is a reliable method for detecting measurable residual disease (MRD) in CBF-AML, with lower limits of detection of <0.01% of transcripts.¹² However, there are reports of CBF translocations being detected during complete remission and a previously published case report describing a patient with detectable fusion in mast cells for 18 months post–hematopoietic cell transplant (HCT),^6,10 raising the question of whether these patients should be considered MRD-positive when they seem otherwise disease-free. Using flow cytometric cell sorting and fluorescent in situ hybridization (FISH), we investigated the source of persistent CBF fusion detection.

We identified 15 bone marrow specimens from 13 patients with CBF fusion detected by RT-PCR but no evidence of MRD by difference from normal multidimensional flow cytometry (ΔN). Eight specimens demonstrated RUNX1::RUNX1T1, and 7 demonstrated CBFB::MYH11 (Table 1, patients 1-13). During ΔN flow cytometric analysis, a noticeable population of mast cells was identified for many patients (mean: 0.35%, range 0.01%-2.5%). These mast cells (bright CD117⁺/CD34^–),¹³ myeloblasts (CD34⁺/CD117 heterogenous), and T cells (CD3⁺) were flow cytometrically sorted (BD FACSAria) for FISH analysis. Mast cells were successfully sorted for all specimens, and myeloblasts (n = 14) and T cells (n = 13) could be sorted for most (Table 1).

When analyzed for the appropriate CBF fusion by FISH, the mast cell population was positive in 13 of 15 specimens while all other cell populations were negative (Table 1). In 2 specimens, the fusion was not detected in any sorted cell population (Table 1, patients 8 and 13). Of the 13 cases with fusion-positive mast cells, 7 (of 8 samples) were positive for RUNX1::RUNX1T1 and 6 (of 7 samples) were positive for CBFB::MYH11. These data suggest that mast cells are the source of positive RT-PCR results in unsorted bone marrow for these specimens, as the fusion was not detected in any cells in the other sorted cell populations (typically 200 cells per population) (Table 1). Mast cells are long-lived leukocytes, activated by immunoglobulin E (IgE) binding, that predominantly reside in vascularized tissues.¹⁴ These data suggest that the longevity and terminal differentiation of mast cells result in a reservoir for CBF fusions, leading to positive RT-PCR results in unsorted specimens.

To confirm our findings, we identified 13 bone marrow specimens (from 6 patients) with known RUNX1::RUNX1T1 fusions at diagnosis that were MRD-negative by ΔN flow cytometry and had detectable mast cells (mean: 0.78%, range: 0.01%-2.5%) (Table 1, patients 14-19). Mast cells were sorted for all specimens along with myeloblasts (n = 12), T cells (n = 12), and neutrophils (side scatter gate [SSC]/CD33⁺, n = 4), then evaluated for RUNX1::RUNX1T1 by FISH (average: 166, range: 42-200 cells). All specimens demonstrated the fusion by FISH in the mast cells and no evidence of fusion in the other sorted cell populations, confirming that mast cells can harbor CBF fusions in the absence of a detectable leukemic population.

Mastocytosis is a clonal neoplastic increase in mast cells, which occurs in either peripheral tissues (cutaneous mastocytosis) or the blood, bone marrow, and/or other extracutaneous organs (systemic mastocytosis [SM]).¹⁵ Despite the known higher frequency of cKIT mutations, SM has rarely been described in association with t(8;21) AML. When evaluating for AML MRD, CD2, CD25, and/or cKIT mutations were evaluated if there were clinical indications of mast cell abnormalities. Although expression of CD2 and CD25 on mast cells fits 1 criteria for SM,^6,7,9,14-16 CD2 and/or CD25 expression on mast cells has also been described after chemotherapy.¹⁷ Of the 19 patients, 2 had aberrant CD2 and/or CD25 expression on mast cells by ΔN flow cytometry (Figure 1A and Table 1, patients 1 and 19). Seven of 19 patients (10 specimens) were evaluated for cKIT mutations (Table 1). Three of 7 patients tested positive for cKIT mutations: 2 on unsorted marrow and 1 in sorted mast cells (sorted myeloblasts from this patient were negative for cKIT mutations). This latter patient, also described above, had CD2 and CD25 aberrantly expressed on the mast cells. Together, these data show that in most patients, mast cells harboring CBF fusions appear normal (phenotypically and by cKIT mutation status) but can be abnormal in a subset of cases. Further studies are needed to determine any clinical implications of abnormal mast cells and if these patients meet other criteria for SM, or if abnormal mast cells in CBF represent unique pathology.

Five patients had multiple specimens evaluated at different time points (Figure 1B and Table 1, patients 9, 10, 14, 18, and 19). Three of these patients (2 with CBFB::MYH11 fusion and 1 with RUNX1::RUNX1T1 fusion) had 2 specimens evaluated over a month apart with persistence of fusion positivity in mast cells only. The other 2 patients had serial specimens evaluated at 4 different time points. One, the patient with a cKIT mutation and aberrant expression of CD2 and CD25 on mast cells, had fusion-positive mast cells that persisted for 8 months. The other patient, who had previously undergone HCT for AML, had detectable fusion positive mast cells over the course of 13 months. All patients had no evidence of MRD by ΔN flow cytometry over the course of testing, and all other sorted cell populations were negative by FISH. Across the entire cohort, fusion-positive mast cells were detected in the absence of MRD by ΔN flow cytometry for an average of 92 days (range: 0-413) since first known negative AML MRD ΔN flow cytometry result (Table 1). Together, these patients demonstrate the longevity of CBF fusion-positive mast cells, which can circulate for prolonged periods, even after successful HCT. This could have significant clinical implications; if this patient had been assessed pre-transplant with RT-PCR on unsorted specimen alone, the patient may not have been considered an HCT candidate due to the detection of the CBF fusion, even though it was confined to terminally differentiated cells. These data show that fusion positivity restricted to mast cells should be considered a separate entity from molecular MRD positivity, and successful transplant can be achieved in these patients, consistent with the previously published case report.⁶ 

These data demonstrate that fusions in CBF-AMLs (RUNX1::RUNX1T1 or CBFB::MYH11) can be detected in the mast cells of disease-free patients, including post-HCT, but are not detected in other sorted cell populations. As terminally differentiated cells, these mast cells do not appear to contribute to disease, suggesting that these patients should be considered MRD-negative. Although RT-PCR is a sensitive and specific assay for monitoring disease state in patients with known fusions, these data demonstrate why caution is required for patients with CBF-AML since these fusions can be retained in the mast cells for months to years. Further studies should be conducted to determine any clinical implications of phenotypically normal vs abnormal mast cells, if CBF fusion-harboring mast cells can impact patient outcome, or if this phenomenon is observed with other genetic abnormalities. These data highlight the importance of combining multiple testing modalities to comprehensively and accurately monitor patients with AML, and the importance of considering the biology of specific AML subtypes.

All specimens were deidentified and the study was performed in accordance with a review exemption issued by the Western Institutional Review Board (Olympia, WA).