Down-Regulation of CD25 Antigen in Hairy Cell Leukemia Patients after Treatment

Background:

Hairy Cell Leukemia (HCL) is a rare hematological malignancy, comprising only of 2% of all leukemias, with an estimated 900 new cases diagnosed each year in the United States. HCL displays a characteristic immunophenotypic profile that include pan-B cell markers including CD103, CD11c, and CD25. World Health Organization guidelines defines two forms of HCL, classic HCL (cHCL) and variant HCL (HCLv) as two distinct clinical entities. Patients with cHCL have a distinct immunophenotypic profile on their malignant leukemic cells including CD20⁺, CD19⁺, CD11c⁺, CD25⁺, CD103⁺, and CD123⁺, while the leukemic cells from patients with HCLv show CD11c⁺, CD20⁺ and CD19⁺, while lacking CD25 and CD123 expression. Some patients with cHCL will retain CD25 positivity while demonstrating negativity for other typical markers, herein termed atypical HCL (aHCL). Presence or absence of CD25 is an important determinant in classifying patients into cHCL and HCLv. Although it has previously been reported that CD25 expression may be lost during treatment with the targeted agent vemurafenib, we sought to identify whether this immunophenotypic change occurs following other treatment types, including standard purine nucleoside analog therapy and with targeted BTK inhibition.

Methods:

Adult patients (≥18 years) with a diagnosis of HCL whom had immunophenotype data collected before and after treatment between 2010 and 2018 were included in the study. Immunophenotype and morphological characteristics of initial and follow-up peripheral blood, bone marrow aspirate, and core biopsy specimens were reviewed and correlated with the treatment received.

Results:

We evaluated 30 HCL patients who underwent different therapies. All available specimens were reviewed and showed morphologic features characteristic for cHCL (n=26, 86.7%), and aHCL (n=4, 13.3%). The median age at HCL diagnosis was 50 years (44-76 years) with male predominance (76%). Patients with aHCL were treated with ibrutinib (n=2) and pentostatin (n=2). Patients with cHCL were treated with pentostatin (n=12), ibrutinib (n=8), vemurafenib (n=4), dabrafenib (n=1), and cladribine (n=1). Bone marrow analyses showed that all the patients had leukemic B-lymphocyte co-expression of CD19, CD20, CD103, CD11c, CD25, and CD123 prior to treatment. Some patients also had a smaller percentage of lymphocytes lacking CD25 expression along with the CD25 positive lymphocytes. Follow-up bone marrow and peripheral blood analysis showed that almost half (n=14, 46%) of treated patients had a partial or complete loss of CD25 expression regardless of the treatment type. Leukemic cells continued to express other HCL signature markers.

Conclusion:

Our study indicates that during the course of disease some patients display a loss of CD25 expression after therapy. This phenomenon was observed across different therapies and is not specific to the type of treatment. This is the first study to show treatment-dependent CD25 variability with pentostatin, ibrutinib and dabrafenib. Our results advocate for caution when using CD25 for the differential diagnosis of cHCL versus HCLv in treated patients. Future studies are needed in larger patient cohorts to determine the overall role and utility of CD25 in the diagnosis of cHCL and HCLv.