Hairy cell leukaemia with low CD103 expression: A rare but important diagnostic pitfall

Abstract

Classical hairy cell leukaemia (HCLc) is an uncommon mature B-cell lymphoproliferative neoplasm (MBCLN) comprising about 2% of lymphoid leukaemias. HCLc is listed as a distinct entity in the World Health Organization (WHO) classification of Tumours of Haematopoietic and Lymphoid Tissues.1 It primarily affects elderly men (median age about 60 years) with a male: female ratio of 4-5:1. Clinical presentation is usually characterized by splenomegaly, pancytopenia and monocytopenia, in the absence of lymph node enlargement. The distinction between HCLc, variant hairy cell leukaemia (HCLv) and other MBCLN is clinically important because patients with HCLc are sensitive to purine analogues such as cladribine (2-chlorodeoxyadenosine) and pentostatin (2'-deoxycoformycin), as well as to interferon-alfa.2, 3 The diagnosis of HCL is based mainly on peripheral blood (PB) and bone marrow (BM) morphology and flow cytometric or immunohistochemical phenotyping of leukaemic cells. On flow cytometry, HCL cells express brightly mature B-lymphoid cell antigens as CD19, CD20, CD22, CD200, are typically positive for CD11c, CD103, CD123 and CD25, while are usually negative or dim for CD5, CD23, CD10, CD79b and CD27. This immunophenotypic pattern is widely used to discriminate HCLc from the morphologically similar HCLv (usually CD103+ but CD25-) and splenic marginal zone lymphoma (SMZL) (usually CD103- and CD11c-).4-7 In particular, CD103 (the αE subunit of integrin αEβ7, also known as the human mucosa lymphocyte antigen) is physiologically expressed by human intestinal intraepithelial T lymphocytes,8 and CD103 expression by B-cell derived HCL and HCLv is considered aberrant. In the largest HCL patient series analysed by flow cytometry, CD103 positivity was documented in all clonal cells of all 169 HCLc and 35 HCLv tested cases (100%), at variance from SMZL.9 Rare CD103-negative HCL cases or with CD103 expression only in a low proportion of leukaemic cells have been described in three single patient reports and in a relatively small series reporting lack of CD103 in 2/34 cases (6%).10-13 However, only one of these five CD103-negative HCL cases was confirmed to express annexin-A1,12 a highly sensitive and specific marker of HCL,14 and only one case was investigated for the BRAF-V600E mutation,13 the genetic hallmark of HCL15 including cases with an aberrant phenotype.16 Here, we reassess the issue of CD103 partial expression in HCLc by analysing a relatively large patient series (n = 71) and document that 3/71 cases (4%) indeed expressed CD103 in a minority of leukaemic cells, while showing typical clinical, immunophenotypic and genetic features of HCLc (including a positivity for annexin-A1 and BRAF-V600E). Between March 2002 and October 2021, we analysed by flow cytometry 35 PB and 36 BM samples from 71 consecutive patients affected by HCLc seen at the Haematology Centre of the “Sapienza” University of Rome. Diagnosis was made according to the current WHO classification criteria.1 All patients provided informed consent for the investigation of both PB and BM materials. They included 55 males and 16 females with an age ranging between 33 and 77 years (median 54). PB and/or BM samples were collected in EDTA. For each tube, 5 × 104 or 10 × 104 total leukocytes were analysed prior to anti-leukaemic treatment. Cells were incubated with an appropriate volume of antibodies in different combinations and multiparameter flow cytometry using 4-6-8 colour combinations was performed with the FACSCalibur or FACSCanto I and II instruments (Becton Dickinson) followed by specific gating strategies with the PAINT-A-GATE and FACSDIVA software (Becton Dickinson). Monoclonal antibodies were used against the following antigens: CD3, CD5, CD22, CD19, CD20, CD10, CD200, CD123, CD103, CD25, CD11c, CD49d, CD38, CD45, CD79-b, FMC-7 (Becton Dickinson), Ig λ and Ig k (Dako). The presence of pathological cells was reported as the percentage of total leukocytes. Immunohistochemistry was performed with a Leica Bond Automated Stainer, using antibodies against BRAF-V600E (clone VE.1; Ventana, Roche), CD103 and annexin-A1 (Leica). Droplet digital PCR for the BRAF-V600E mutation was carried out on the Bio-Rad platform using genomic DNA and the validated Bio-Rad droplet digital PCR (ddPCR) mutation detection assay catalogue# 10049551. All 71 HCL cases showed surface immunoglobulin light chain restriction (33 Igλ and 38 IgK); they were positive for CD19, CD20, CD22, CD200, CD79b, FMC7, CD11c and CD25, and were negative for CD5 and CD10. CD38 was positive only in 4/61 cases (7%). PB samples showed a median HCL cells percentage of 10% (range: 0.1%–80%) compared to a median of 14.5% (range: 3%–87%) on BM samples. Three cases (3/71, 4%) - despite expressing CD25, CD11c, and CD200 on all clonal cells (which were 15%, 22% and 37% of total leukocytes, respectively) - had CD103 expressed only on a fraction of leukaemic cells (40%, 27%, and 10.8%, corresponding to 5%, 6% and 4% of total leukocytes) (Figure 1A). In all three cases, a low CD103 expression was confirmed on immunohistochemistry of BM biopsies (Figure 1B), while both a clonal BRAF-V600E mutation and annexin-A1 expression were documented by immunohistochemistry (Figures 1C–D) and/or ddPCR (Figure 1E). The clinical picture of the three cases, featuring cytopenias (including monocytopenia), splenomegaly and circulating leukaemic cells with hairy morphology (not shown), was typical of HCL. The three patients also responded to therapies known to be effective in HCL, that is, cladribine (in 2 cases) and interferon-alfa (in 1 patient who refused chemotherapy) that led to a complete responses currently ongoing at 4, 7 and 11 years of follow-up, respectively (Table 1). In conclusion, although CD103 is a robust HCL marker, it can occasionally be expressed in only a minority of the leukaemic clone, a pitfall that should not lead to automatically exclude a diagnosis of HCL, but rather to perform a thorough phenotypic and genetic work-up in order to reach a correct diagnosis and to select the appropriate treatment. This work was supported by Associazione Italiana Ricerca sul Cancro (AIRC), Special 5 × 1000 Program Metastases (21198), Milan (Italy) to RF and BF. SI was supported by ROMAIL ONLUS. MSDP, RF, ET and AG designed the research study and wrote the manuscript; MSDP, SI, MGN, AP,2 AS and NP carried out the laboratory work; MSDP, PM, MC and MLDL analysed the data; AP,1 GMD, IDG collected patients' data; RF, ET, MM, AP1 and BF revised the manuscript.

Topics & Concepts