Aberrant Expression of CD123 (Interleukin-3 Receptor-Α) on Neoplastic Mast Cells

Letters to the Editor 1605 Aberrant expression of CD123 (interleukin-3 receptor-α) on neoplastic mast cells

Leukemia (2015) 29, 1605–1608; doi:10.1038/leu.2015.16 incubated at room temperature for 20 min in the dark and then washed again. Surface antigen expression was interrogated using the 4-color multiparametric flow cytometer, FACSCalibur (BD The high-affinity receptor for interleukin-3 (IL-3) comprises the Biosciences). Data analysis was performed using CellQuest Pro ligand-binding α-subunit (CD123) plus the primary signaling Software (BD Biosciences). Formalin-fixed and decalcified paraffin- common β (βc) subunit (CD131), the latter also being shared embedded BM biopsies were sectioned into 4 μmslicesandaffixed with receptors for IL-5 and granulocyte-macrophage colony on glass slides. Samples were heated for half an hour at 56 °C, and stimulating factor (GM-CSF).1 IL-3 is a polyfunctional CSF that acts then deparaffinized in xylene, rehydrated in a graded alcohol series primarily on committed granulocyte-macrophage progenitor cells and washed in water. Heat-induced epitope retrieval was used for and their immediate progeny; in vivo administration of IL-3 in both CD123 and CD117/KIT antigens. Enzyme pretreatment epitope normal adult mice leads to increased levels of circulating retrieval was used for Tryptase antigen. Endogenous peroxidase eosinophils, granulocytes and monocytes as well as increased activity was quenched in a bath of methanol and hydrogen splenic mast cells.2 CD123 is expressed in a variety of peroxide. We stained samples for Tryptase (Dako, Carpinteria, CA, hematological neoplasms including blastic plasmacytoid dendritic USA) at 1:1500, CD123 (clone 7G3, BD Pharmingen, Franklin Lakes, cell neoplasm (BPDCN), acute myeloid leukemia (AML), acute NJ, USA) at 1:2000 and CD117/KIT (Dako) at 1:500. Samples were lymphoblastic leukemia (ALL), chronic myeloid leukemia (CML), incubated overnight with antibodies and peroxidase activity was Hodgkin’s lymphoma and hairy cell leukemia (reviewed by Testa localized for all samples with 3,3′-diaminobenzidine using Ventana et al.3). In normal bone marrow (BM), o1% of CD34+/CD38 − staining and detection platforms (Ventana Medical Systems, Inc. stem/progenitor cells express CD123; in contrast, 98% of Tucson, AZ, USA). Hematoxylin and eosin counterstaining was done corresponding cells from AML patients show strong expression per standard methods. of this marker.4 Similarly, CD123 expression has also been Mast cell immunophenotyping by FC was performed on 14 SM demonstrated in CD34+/CD38 − cells in the majority of patients patients; data were informative for 6 patients (8 samples) for with myelodysplastic syndromes (MDS) and CML.5,6 In most AML whom an adequate number of mast cells were identified patients, bulk CD34+ leukemic blasts virtually uniformly express for immunophenotyping: (1) indolent SM (ISM, n =2, PB=1, CD123, and the putative leukemic stem cell (CD34+/CD38 − / paired PB/BM = 1); (2) aggressive SM (ASM, n = 1, paired PB/BM = 1) CD123+) has been shown to successfully engraft immune- and (3) SM with an associated hematological neoplasm (SM-AHD, deficient NOD/SCID mice. Early-phase therapeutic studies n = 3, PB (myelodysplastic syndrome/myeloproliferative neoplasm- targeting CD123 in AML patients are currently underway.7,8 There not otherwise specified, (MDS/MPN-NOS) = 2, BM (polycythemia is also ubiquitous, strong expression of functionally competent vera) = 1). Clinical and laboratory characteristics of these patients CD123 on CD4+/CD56+ blasts from BPDCN patients, an are described in Supplementary Table 1. A median total 653 625 observation that has been successfully exploited in preliminary cells per sample were interrogated (range 1 72 050–2 513 850). studies using an immunotoxin to target CD123 in these patients.9 Neoplastic mast cells were defined as CD117++/SSC++/CD45+ Whereas CD123 is not expressed in normal or reactive mast cells,10 +/CD34 − /CD25+/FcεRI+ (Figure 1). The majority of mast cells for CD34+/CD38 − stem/progenitor cells from systemic mastocytosis five patients expressed CD123 (no. 1–5, CD123 percentage (SM) patients reportedly express this marker.5 However, there are positive, PB/BM): no. 1, CD123 = 91% (PB)/ BM not done; no. 2, conflicting data regarding CD123 expression on neoplastic mast CD123 = 76% (PB)/ 82% (BM); no. 3, CD123 = 85% (PB)/ BM not cells from SM patients as detected by flow cytometry (FC) versus done; no. 4, CD123 = 88% (PB)/ BM not done; and no. 5, immunohistochemistry (IHC).11,12 In the current study, we CD123 = 84% (BM)/PB not done. In contrast, one patient (no. 6) analyzed, using both FC and IHC techniques, as to whether showed a minority of mast cells to be CD123 positive: (33% PB and neoplastic mast cells in peripheral blood (PB) and BM from SM 28% BM). For both SM patients who had paired PB and BM patients express CD123 in order to build a rationale for use of samples analyzed, CD123 expression was concordant in the two CD123-targeting therapies in this patient population. samples. FcεRI expression was brighter on mast cells from BM as The current study was approved by our institutional review compared with PB (data not shown). board. All patients provided written informed consent for PB and/or CD123 immunophenotyping by IHC was performed on 3 control BM sample collection. Consecutive patients with confirmed SM samples (patients with clinically suspected SM, but with a negative were recruited from the Mastocytosis Clinic in the Division of workup) and 5 additional SM patients: (1) ISM, n = 3; (2) ASM, n =1; Hematology; no other selection criteria were applied. The diagnosis and (3) SM-AHD (MDS/MPN-NOS), n = 1 (Supplementary Table 1). of SM and its subclassification was according to WHO (World Health In control BM sections, rare scattered interstitially distributed cells Organization) criteria.13 For FC, PB and BM samples were subjected showed strong CD123 staining, likely representing plasmacytoid to ammonium chloride lysis. Cells were then washed twice in dendritic cells. Weak staining of vascular endothelial cells in some staining buffer (phosphate-buffered saline containing 1 mM EDTA venules/capillaries was observed; there was no readily perceptible and 0.5% bovine serum albumin), resuspended in 100 μland staining of other hematopoietic cells. For the SM cases, CD123 appropriate dilution of corresponding fluorophore-conjugated staining was assessed on recognizable tight aggregates of mast surface antibody was added. Primary mouse anti-human antibodies cells, evident by their tryptase and CD117 staining characteristics. used were as follows: PE-CD123 (cat. no. 340545), FITC-CD25 In the two cases with ASM and SM-AHD, morphologically atypical (cat. no. 555431), PerCP-CD45 (cat. no. 340665), FITC-CD34 and tryptase-positive mast cells showed a strong, uniform CD123 (cat. no.555821), APC-CD117 (cat. no. 341096) (BD Biosciences, staining pattern (Figures 2e and f). For the three ISM cases, San Jose, CA, USA) and PE mouse anti-human FcεRI (cat. no. mast cell CD123 staining appeared more variable relative to SM2257R) (Acris Antibody, San Diego, CA, USA). Samples were the former (Figures 2a–d). In one ISM case, a focal increase of

Accepted article preview online 2 February 2015; advance online publication, 17 February 2015

Figure 1. Flow cytometry plots to highlight immunophenotype of neoplastic mast cells from a patient with systemic mastocytosis. (a) Gating strategy to identify circulating mast cells: CD117+, high side scatter (SSC); (b) gated cells exhibit bright CD45 expression; (c–f) expression of CD25, FcεRI, CD34 and CD123 on gated cells (green open line) as compared with isotype control (ﬁlled red line). Neoplastic mast cell phenotype is CD117++/CD45++/SSC++/CD34 − /CD25+/ FcεRI+; the majority of these cells express CD123.

plasmacytoid dendritic cells (strongly CD123 positive) was approaches that are rapidly evolving, given aberrant expression of seen in close proximity of the mast cell cluster/infiltrate this marker on both stem/progenitor (CD34+/CD38 − ) cells as well (Figures 2c and d). as bulk tumor (mast) cells in this disease. These observations are Our data, using two methods, namely FC and IHC, identified clinically relevant given the current unmet need for effective aberrant expression of CD123 on neoplastic mast cells in the therapies for aggressive or advanced SM;14 although both majority of patients with SM. As FC and IHC were not performed cladribine and interferon-α have modest efficacy in this setting, on the same patient set, data from the two methods in the current they have limitations in terms of transient treatment responses study are not directly comparable. Furthermore, variability in and dose-limiting side effects. Furthermore, the role of tyrosine staining between methods may reflect epitope differences for kinase inhibitors in SM therapy remains unclear; although recent antibodies from different clones. Overall, however, our data data with midostaurin have been encouraging in this regard,15 in combination with previous observations suggest that SM additional information with longer follow-up is awaited. The presents a unique opportunity for CD123-targeting treatment limitations of this study include its relatively small sample size; in

Figure 2. Immunohistochemical stain highlighting CD123 expression on bone marrow mast cells from patients with systemic mastocytosis. Three representative cases are shown. Each row represents a single patient and shows representative photomicrographs of CD123 immunohistochemical stain (left panels: a, c, e) and CD117 (c-KIT) immunohistochemical stain (right panels: b, d, f) on the core biopsy. (a, b) and (c, d) Two patients with indolent systemic mastocytosis, showing variability in mast cell CD123 staining. (a) Relatively weak CD123 staining within mast cell cluster is shown. The interspersed cells with strong CD123 staining represent plasmacytoid dendritic cells (PDCs). (b) Stronger mast cell CD123 staining with clustering of PDCs in close proximity of the mast cell cluster is shown. (e, f) Shown is a patient with aggressive systemic mastocytosis, showing strong uniform mast cell CD123 staining (e). addition, CD123 expression on non-mast lineage cells in SM-AHD ACKNOWLEDGEMENTS patients was not systematically studied. A larger study is currently Intramural funds from the Department of Laboratory Medicine and Pathology underway at our institution to confirm and extend the aforemen- supported the immunohistochemistry experiments. tioned findings. AUTHOR CONTRIBUTIONS AP, ER and CB designed the study. AP, TL and AT analyzed the data and wrote CONFLICT OF INTEREST the first draft of the manuscript. TKK helped set up the flow cytometry experiments and analyze the data. TL and CF performed the flow cytometry AP received research funding from Stemline Therapeutics Inc. for the flow cytometry experiments. AP is Principal Investigator for a forthcoming clinical trial of SL-401 experiments. DZ and MMP contributed patients/patient samples to the study. for treatment of myeloproliferative neoplasms. SL-401 is a CD123-targeting DC, KKR and CAH performed the immunohistochemistry experiments and/or investigational drug manufactured by Stemline Therapeutics Inc. CB and ER are reviewed the bone marrow slides. All authors reviewed, provided input and employees of Stemline Therapeutics Inc. approved the final draft of the manuscript.

© 2015 Macmillan Publishers Limited Leukemia (2015) 1600 – 1618 Letters to the Editor 1608 1 1 2 1 1 A Pardanani , T Lasho , D Chen , TK Kimlinger , C Finke , 7 He SZ, Busfield S, Ritchie DS, Hertzberg MS, Durrant S, Lewis ID et al. A phase 1 D Zblewski1, MM Patnaik1, KK Reichard2, E Rowinsky3, CA Hanson2, study of the safety, pharmacokinetics, and anti-leukemic activity of the C Brooks3 and A Tefferi1 anti-CD123 monoclonal antibody, CSL360, in relapsed, refractory or high-risk 1Division of Hematology and Department of Medicine, Rochester, acute myeloid leukemia (AML). Leuk Lymphoma 2014; e-pub ahead of print 20 MN, USA; November 2014; doi:10.3109/10428194.2014.956316. 2Division of Hematopathology and Department of Laboratory 8 Frankel AE, Konopleva M, Hogge D, Rizzieri D, Brooks C, Cirrito T et al. Medicine and Pathology, Mayo Clinic, Rochester, MN, USA and Activity and tolerability of SL-401, a targeted therapy directed to the interleukin-3 3Stemline Therapeutics Inc., New York, NY, USA receptor on cancer stem cells and tumor bulk, as a single agent in patients with advanced hematologic malignancies. J Clin Oncol 2013; 31:15. E-mail: [email protected] 9 Frankel AE, Woo JH, Ahn C, Pemmaraju N, Medeiros BC, Carraway HE et al. Activity of SL-401, a targeted therapy directed to interleukin-3 receptor, in blastic plas- 124 – REFERENCES macytoid dendritic cell neoplasm patients. Blood 2014; :385 392. 10 Valent P, Besemer J, Sillaber C, Butterfield JH, Eher R, Majdic O et al. Failure to 1 Woodcock JM, Bagley CJ, Zacharakis B, Lopez AF. A single tyrosine residue in the detect IL-3-binding sites on human mast cells. J Immunol 1990; 145: 3432–3437. membrane-proximal domain of the granulocyte-macrophage colony-stimulating 11 Teodosio C, Garcia-Montero AC, Jara-Acevedo M, Sanchez-Munoz L, Alvarez-Twose I, factor, interleukin (IL)-3, and IL-5 receptor common beta-chain is necessary Nunez R et al. Mast cells from different molecular and prognostic subtypes of and sufficient for high affinity binding and signaling by all three ligands. systemic mastocytosis display distinct immunophenotypes. J Allergy Clin Immunol J Biol Chem 1996; 271: 25999–26006. 2010; 125:719–726. 2 Metcalf D, Begley CG, Johnson GR, Nicola NA, Lopez AF, Williamson DJ. Effects 12 Moonim MT, Kossier T, van Der Walt J, Wilkins B, Harrison CN, Radia DH. of purified bacterially synthesized murine multi-CSF (IL-3) on hematopoiesis in CD30/CD123 expression in systemic mastocytosis does not correlate with normal adult mice. Blood 1986; 68:46–57. aggressive disease. Blood 2012; 120:21. 3 Testa U, Pelosi E, Frankel A. CD 123 is a membrane biomarker and a therapeutic target in hematologic malignancies. Biomark Res 2014; 2:4. 13 Horny HP, Metcalfe DD, Bennett JM, Bain BJ, Akin C, Escribano L et al. 4 Jordan CT, Upchurch D, Szilvassy SJ, Guzman ML, Howard DS, Pettigrew AL et al. Mastocytosis. In: Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H et al. fi The interleukin-3 receptor alpha chain is a unique marker for human acute (eds) WHO Classi cation of Tumors of Hematopoietic and Lymphoid Tissues, myelogenous leukemia stem cells. Leukemia 2000; 14: 1777–1784. 4th edn. International Agency for Research and Cancer (IARC): Lyon, 2008; pp – 5 Florian S, Sonneck K, Hauswirth AW, Krauth MT, Schernthaner GH, Sperr WR et al. 54 63. Detection of molecular targets on the surface of CD34+/CD38— stem cells in 14 Pardanani A. Systemic mastocytosis in adults: 2013 update on diagnosis, risk various myeloid malignancies. Leuk Lymphoma 2006; 47:207–222. stratification, and management. Am J Hematol 2013; 88:612–624. 6 Frolova O, Benito J, Brooks C, Wang RY, Korchin B, Rowinsky EK et al. SL-401 and 15 Gotlib J, Kluin-Nelemans HC, George TI, Akin C, Sotlar K, Hermine O et al. SL-501, targeted therapeutics directed at the interleukin-3 receptor, inhibit the Durable responses and improved quality of life with midostaurin (PKC412) in growth of leukaemic cells and stem cells in advanced phase chronic myeloid advanced systemic mastocytosis (SM): updated stage 1 results of the Global leukaemia. Br J Haematol 2014; 166:862–874. D2201 trial. Blood 2013; 122:21.

Supplementary Information accompanies this paper on the Leukemia website (http://www.nature.com/leu)

OPEN

Quantifying ultra-rare pre-leukemic clones via targeted error-corrected sequencing

Leukemia (2015) 29, 1608–1611; doi:10.1038/leu.2015.17 mutations. With this method, detecting mutations at 2–5% variant allele fraction (VAF) requires costly and time-intensive deep The quantification of rare clonal and subclonal populations from a resequencing and identifying lower frequency variants is impractical heterogeneous DNA sample has multiple clinical and research regardless of sequencing depth. Recently, various methods have applications for the study and treatment of leukemia. Specifically, been developed to circumvent the error rate of NGS.4,5 These in the hematopoietic compartment, recent reports demonstrate the methods tag individual DNA molecules with unique oligonucleotide presence of subclonal variation in normal and malignant indexes, which enable error correction after sequencing. hematopoiesis,1,2 and leukemia is now recognized as an oligoclonal Here we present a direct application of error-corrected disease.3 Currently, clonal heterogeneity in leukemia is studied sequencing (ECS) to study clonal heterogeneity during leukemo- using next-generation sequencing (NGS) targeting subclone-specific genesis and validate the accuracy of this method with a series

Figure 1. Benchmarking for ECS and the identification of rare pre-leukemic mutations. (a, b) DNA extracted from a diagnostic leukemia sample with known mutations in RUNX1 (a) and IDH2 (b) was serially diluted into non-cancer, unrelated human DNA. Two replicates were run per sample/dilution. The coefficient of determination (r2) between diluted tumor concentration in the sample and VAF in the generated read families was 0.9999 and 0.9991 for RUNX1 and IDH2, respectively. (c) The VAF at every nucleotide not expected to contain mutations in the dilution series experiment were analyzed to determine the error profile of the error-corrected consensus sequences compared with conventional deep sequencing. A cumulative distribution function of VAF demonstrated a reduced error profile in read families relative to conventional deep sequenced reads. (d) The most frequent class of substitution seen in read families was in G to T (C to A) transversions, which was consistent with oxidative conversion of guanine to 8-oxo-guanine. (e, f) The leukemia-specific variants identified in ASXL1 and U2AF1 at diagnosis (circled) were not distinguishable from sequencing errors in the same substitution class by conventional deep sequencing. (g, h) Targeted error-corrected sequencing identified the ASXL1 variant in the 2002 banked sample at 0.004 VAF and the U2AF1 variant in the 2004 banked sample at 0.009 VAF.

Accepted article preview online 3 February 2015; advance online publication, 20 February 2015