Acute myeloid leukaemia (AML) is a heterogeneous haematologic malignancy characterised by dysfunctional proliferation and differentiation of myeloid progenitor cells.1 Though the standard 7+3 treatment regimen, consisting of cytarabine, daunorubicin, and etoposide induces remission, a significant proportion of patients relapse resulting in negative outcomes overall.2
To address the need for new effective agents to treat AML, several immunotherapeutic therapies are being investigated in various AML treatment regimens. CD33 is a myeloid cell surface protein present on leukaemic cells in >85% of AML patients and consists of 2 extracellular domains, the IgV and IgC domains, as well as an intracellular immunoreceptor tyrosine-based inhibition motif (ITIM).3 CD33 internalisation upon antibody binding makes it a lucrative target in AML with treatment strategies including antibody-drug conjugates, such as gemtuzumab ozogamicin (GO), an antibody–drug conjugate reapproved in 2017 for treatment of AML, and other immunotherapeutics.4–6 While CD33-directed therapies have revolutionised AML treatment strategies, studies have shown evidence of interpatient variation in CD33-directed therapeutic response.7 In particular, the authors of this article previously reported occurrence of a splicing single nucleotide polymorphism, rs12459419 (C>T), which resulted in skipping of exon 2 and formation of a shorter CD33 isoform (CD33D2) lacking the IgV domain.8 Given that all currently available CD33-directed therapies and immunophenotyping diagnostics are based on recognition of the IgV domain, the minor allele of rs12459419 is associated with lower cell surface levels of CD33 and altered response to GO. Consequentially, patients with at least one variant allele do not benefit from inclusion of GO in their chemotherapeutic regimen.8 The lack of benefit observed in heterozygous patients remains an enigma and warrants in-depth investigations into CD33D2 biology. Thus, the aim of this study was to develop novel antibodies which recognise CD33D2 and utilise these antibodies to characterise the expression pattern, localisation, and therapeutic implications of CD33D2, including its potential as a novel drug target and role in the compromised response to GO observed in heterozygous patients.
Novel CD33 antibodies were developed by immunising Balb/c or C57BL/6 mice with peptides spanning the IgC domain. The novel antibodies, 5C11-2 and HL2541, were then confirmed for recognition of CD33 by Western blotting using the CD33-Ba/F3 cell line engineered to express either CD33FL or CD33D2. Subsequently, immunofluorescence assays using the novel antibodies and the IgV targeting antibody, P67.6, were performed to assess the expression pattern and localisation of CD33 isoforms using the Ba/F3 cell line system, AML cell lines of different rs12459419 genotypes, and primary AML bone marrow specimens.
Western blotting analysis using the novel IgC antibodies and the ITIM domain-directed antibody, E6, as a control, confirmed recognition of both CD33FL and CD33D2 by the novel IgC-directed antibodies; however, only CD33D2 was recognised on the cell surface in flow cytometric assays. These results indicate a potential steric hindrance caused the clustered extracellular Ig-folds which may stifle recognition of CD33FL by IgC-targeting antibodies. Flow cytometry performed on AML cell lines stained with P67.6, 5C11-2, and HL2541 revealed extracellular and intracellular recognition CD33FL while CD33D2 was only recognised intracellularly regardless of genotype. Ongoing work continues using primary AML specimens to fully elucidate the cell surface expression of CD33D2 and the potential for its targeting. Once confirmed, these results hold potential for establishing CD33 biology and understanding the associated therapeutic implications.