Page 132 - 2021_04-Haematologica-web

W. Wang et al.
Immunophenotypically, BPDCN cells are typically posi- tive for CD4, CD56, CD123, HLA-DR, TCL1, and TCF4, and are negative for lineage-specific antigens for B cells (e.g., CD19), T cells (surface and cytoplasmic CD3) as well as myeloid cells (myeloperoxidase).3,8 Monocytic markers such as CD64 are also negative. Our group has recently demonstrated that co-expression of CD123 and TCF4, as determined by immunohistochemistry constitutes a high- ly reliable marker for BPDCN.9
Despite significant advances in immunophenotypic characterization of BPDCN at baseline, data regarding the distinction between reactive/normal PDC and BPDCN cells remain limited. This limitation raises diagnostic challenges particularly in the evaluation of BM samples with a minimal disease burden, either at presentation in patients with predominantly extramedullary disease or after therapy in patients evaluated for measurable/mini- mal residual disease (MRD). With advances in treatment options for BPDCN patients and the importance of achievement of disease remission for allogeneic stem cell therapy, the need for reliable and reproducible criteria to assess BM samples for potential low-level BPDCN involvement has gained increased attention. An assay that can reliably distinguish neoplastic PDC from back- ground reactive PDC becomes important. Variable num- bers of reactive PDC are detected routinely in the BM by flow cytometry immunophenotyping and/or immunohis- tochemistry. Similar to their neoplastic counterparts, reactive PDC are positive for CD4, CD123, CD303, HLA- DR and TCF4, and they lack expression of lineage-specif- ic antigens. CD56, a marker frequently expressed in BPDCN, is the only marker being used to date to distin- guish neoplastic from reactive PDC. However, CD56 expression can be found in a small subset of reactive PDC.10-12 Thus, distinguishing CD56+ BPDCN from CD56+ reactive PDC becomes quite challenging in the assess- ment of post treatment BM specimens, which often con- tain reactive PDC.
In this study, our aim was to characterize the immunophenotype of BPDCN, with particular focus on the differences between BPDCN and normal/reactive PDC. From understanding these differences, we devel- oped and validated a ten-color clinical-grade flow cytom- etry immunophenotyping panel and compared its per- formance to that of orthogonal tools for residual disease evaluation.
Methods
Study group
We identified all patients with BPDCN diagnosed at The University of Texas MD Anderson Cancer Center between 2010 and 2019. All patients fulfilled the diagnostic criteria for BPDCN as defined in the World Health Classification. Patients for whom flow cytometry immunophenotyping had been performed on BM specimens were included in this study. A control group of patients who had BM evaluation by flow cytometry immunophenotyping were also included to study reactive PDC; this group included patients who underwent BM staging for lymphoma or had hematologic diseases other than BPDCN and were in complete remission with or without stem cell transplan- tation. This study was approved by the University of Texas MD Anderson Cancer Center Institutional Review Board and was conducted in accordance with the Declaration of Helsinki.
Table 1. The list of antibodies used in our flow cytometric panels.
Panel #
Panel #1
Panel #2
Panel #3
Antibody list
Tube 1: CD7/CD33/CD19/CD34/CD13/CD2/CD38/CD45 Tube 2: HLA-DR/CD117/CD4/CD34/CD123/CD38/CD45 Tube 3: CD41/CD36/CD56/CD34/CD64/HLA-DR/CD14/CD45 Tube 4: CD5/CD25/CD22/CD34/CD38/CD15/CD45
Tube 5: TdT/MPO/CD34/CD3/cytoCD3/CD45 (cyto tube)
HLA-DR/CD64/CD4/CD33/CD56/CD45/CD303/CD123
CD2/CD4/CD7/CD38/CD45/CD56/CD64/CD123/CD303/ HLA-DR
Flow cytometric immunophenotypic analysis
BM aspirate specimens were collected in EDTA anticoagulant tubes, and processed within 12 h of collection using a standard lyse/wash technique (PharmLyseTM, BD Biosciences, San Diego, CA, USA). For each analysis a minimum of 200,000 events was acquired on FACSCanto II instruments (8-color and 10-color, BD Biosciences).
At the time of initial diagnosis, a comprehensive panel designed for the workup of acute leukemia was performed routinely (panel #1 in Table 1). This panel included lineage-defining markers for B, T, myeloid, and monocytic cells, as well as markers (CD4, CD123, HLA-DR, CD56) necessary for initial screening of BPDCN. When a diagnosis of BPDCN was suspected from panel #1 analysis, an additional panel (panel #2 in Table 1) was used for further charac- terization and confirmation.
Based on the findings of the current study, a one-tube, ten-color assay (panel #3 in Table 1) was subsequently constructed and val- idated for distinguishing BPDCN cells from reactive PDC.
Immunohistochemistry
Immunohistochemical studies were performed using formalin- fixed, paraffin-embedded BM core biopsy or aspirate clot speci- mens.13 TCF4/CD123 double staining was performed using a pre- viously described protocol.9
Results
Immunophenotype of blastic plasmacytoid dendritic cell neoplasm
A total of 39 patients with a diagnosis of BPDCN were studied, including 30 men and 9 women with a median age of 69 years (range, 3-87 years). Flow cytometry immunophenotyping was performed using panel #1 (Table 1) and the more recent cases were also tested using panel #2 (Table 1). The median number of BPDCN cells detected by the flow cytometry was 18% (range, 0.1- 91%). The immunophenotype of BPDCN in these 39 cases is summarized in Figure 1 and Online Supplementary Table S1.
BPDCN cells were positive for CD45, falling into the “blast” gate on CD45/SSC in all cases (39/39, 100%). CD45 expression was often present at a similar level as, or slightly higher than, that of granulocytes (Figure 2A) with the exception of three cases (8%) which showed signifi- cantly lower CD45 expression (dimmer than granulo- cytes) (Figure 2B). HLA-DR as well as CD123 expression was uniformly positive in all cases. Cells were positive for CD4 in all 38 cases assessed, uniform in 34 (89%) and par- tially in four (11%) cases. BPDCN cells were positive for CD56 in 97% (36/37) of cases, with mostly (92%, 33/36) uniform and occasionally partial (3/36, 8%) expression. The only case in which CD56 was not expressed was a 3-
1048
haematologica | 2021; 106(4)