ALBERT

Description: Background: CD33, the target for the anti-AML immunoconjugate, gemtuzumab ozogamicin (GO; Mylotarg™), contains two cytoplasmic immunoreceptor tyrosine-based inhibitory motifs (ITIMs). We have previously shown that these motifs control uptake of antibody-bound CD33 and GO-induced cytotoxicity. In this study, we determined which phosphorylation state favors uptake of antibody-bound CD33, identified proteins that bind to CD33 in an ITIM-dependent manner, and assessed their importance for CD33 internalization by siRNA-based gene silencing. Methods: Internalization of anti-CD33 antibodies was measured by flow cytometry in the presence or absence of the tyrosine phosphatase inhibitor, pervanadate, in human CD33+ AML cell lines (ML-1, HL-60, NB4, U937, TF-1) and CD33− Jurkat T cells infected with wild-type and mutant CD33. Pull-down experiments were performed with glutathione S-transferase (GST) proteins fused to phosphorylated cytoplasmic tails of CD33, using human myeloid cell lysates. Co-immunoprecipitations were performed with myeloid cell lines expressing HA-tagged wild-type CD33. Lentivirus-based siRNA constructs were generated for gene silencing, and expressed in human CD33+ AML cell lines. Results: Pervanadate significantly increased uptake of anti-CD33 antibodies in human AML cell lines; this effect was dependent upon the integrity of the ITIMs and was prevented by co-treatment with the Src tyrosine kinase inhibitor PP2, suggesting that Src family kinase-dependent phosphorylation of the ITIMs critically controls uptake of antibody-bound CD33, possibly by altering which proteins binds to CD33 or by facilitating binding of adaptor-proteins required for endocytosis. We identified several proteins, including the tyrosine phophatases, SHP-1 and SHP-2, and the non-receptor tyrosine kinase, Syk, which bound to phosphorylated wild-type and mutant CD33 in a manner that paralleled the endocytic properties of the corresponding CD33 protein. Since these three proteins have been implicated in endocytic processes of other cell surface proteins, we assessed their role in uptake of antibody-bound CD33 by siRNA-mediated gene silencing. Simultaneous depletion of SHP-1 and SHP-2, but not SHP-1 or SHP-2 alone, significantly increased internalization of antibody-bound CD33 in the two AML cell lines with the highest cell surface expression of CD33, whereas no effect was seen in two other cell lines with lower CD33 expression levels. In contrast, depletion of Syk, whose expression has previously been correlated to the inhibitory effect of anti-CD33 antibodies on AML cell growth, failed to affect antibody internalization in the cell lines assessed. Conclusion: These studies indicate that the phosphorylation status of the ITIMs controls uptake of antibody-bound CD33. In line with this model, SHP-1 and SHP-2, which have been shown to dephosphorylate CD33 in vitro, can affect this endocytic process. Thus, our data imply manipulation of the phosphorylation state of CD33, e.g. by activating Src family kinases or interfering with phosphatases as a novel means to increase uptake of anti-CD33 antibody-based therapeutics such as GO. Finally, the variable effect of SHP-1 and SHP-2 depletion suggests that there are significant cell-type specific differences in the response to anti-CD33 antibody ligation, for example differences in tyrosine phosphorylation levels and/or activation/recruitment or redundancies of tyrosine phosphatases.

Description: Gemtuzumab ozogamicin (GO; Mylotarg), a novel immunoconjugate used for treatment of acute myeloid leukemia (AML), contains the humanized anti-CD33 antibody (hP67.6) as a carrier to facilitate cellular uptake of the toxic calicheamicin-γ1 derivative. By use of lentivirus-mediated gene transfer to manipulate CD33 expression in myeloid cell lines that normally lack CD33 (murine 32D cells) or have very low levels of CD33 (human OCI-AML3 and KG-1a cells), we here show a quantitative relationship between CD33 expression and GO-induced cytotoxicity. The CD33 cytoplasmic immunoreceptor tyrosine-based inhibitory motifs (ITIMs) control internalization of antibody bound to CD33. Disruption of the ITIMs by introduction of point mutations not only prevented effective internalization of antibody-bound CD33 but also significantly reduced GO-induced cytotoxicity. Together, our data imply a pivotal role of both the number of CD33 molecules expressed on the cell surface and the amount of internalization of CD33 following antibody binding for GO-induced cytotoxicity and suggest novel therapeutic approaches for improvement of clinical outcome of patients treated with GO.

Description: Background: CD33, the target for the anti-leukemia immunoconjugate, gemtuzumab ozogamicin (GO; Mylotarg™), is a transmembrane glycoprotein that contains two cytoplasmic immunoreceptor tyrosine-based inhibitory motifs (ITIMs). Although we have previously shown that disruption of the ITIMs prevents effective uptake of antibody-bound CD33 and significantly reduces GO-induced cytotoxicity, the mechanisms underlying this uptake and intracellular trafficking of the antibody-CD33 complex are not known. In this study, we tested whether CD33 is a target for monoubiquitination, a posttranslational modification that marks proteins for lysosomal degradation, and how this modification relates to tyrosine phosphorylation and endocytosis. Methods: CD33− Jurkat and 32D cells transduced with a lentiviral vector expressing either wild-type or mutant CD33 were analyzed for CD33 expression as well as internalization of anti-CD33 antibody by flow cytometry. Pull-down experiments were performed with glutathione S-transferase (GST) proteins fused to phosphorylated cytoplasmic tails of CD33 or TKB domain of Cbl/Cbl-b, using human myeloid cell lysates. For co-immunoprecipitation experiments, constructs encoding wild-type and mutant CD33, ubiquitin, Cbl/Cbl-b, and wild-type Fyn were transfected into HEK293T cells. Results: In engineered Jurkat cells, treatment with either anti-CD33 antibody or the tyrosine phosphatase inhibitor pervandate increased tyrosine phosphorylation of CD33. Pervanadate enhanced uptake of antibody-bound CD33; this effect was dependent upon the integrity of the ITIMs and was prevented by co-treatment with the Src tyrosine kinase inhibitor PP2. CD33 interacted with the TKB domains of Cbl and Cbl-b in GST fusion protein pulldown assays. Similarly, Cbl and Cbl-b were co-immunoprecipitated with CD33 in transfected 293T cells. Experiments in 293T cells further showed that CD33 is monoubiquitinated in an ITIM-dependent manner. Cbl or Cbl-b significantly increased the amount of CD33-associated ubiquitination; Cbl/Cbl-b-dependent ubiquitination could further be enhanced by co-transfected Fyn. Finally, 32D cells expressing lysine-to-arginine mutants of CD33 displayed much higher levels of surface CD33 but had reduced internalization of antibody-bound CD33 compared to cells expressing wild-type CD33, consistent with a reduced internalization/degradation of the lysine-to-arginine mutants. Conclusion: These studies indicate that Src-family kinase dependent phosphorylation favors internalization of antibody-bound CD33 and identify Cbl family proteins as potential binding partners of CD33. Importantly, CD33 is a target for ITIM-dependent monoubiquitination, and Cbl family proteins can act as an E3 ligase in this reaction. Our data therefore suggest a model where lysosomal routing and degradation of antibody-bound CD33 is secondary to Src-family kinase-induced phosphorylation of CD33 with subsequent phosphotyrosine-dependent binding of Cbl to and monoubiquitination of CD33.